BMJ - 25 May 2013

THIS WEEK
BMJ | 25 MAY 2013 | VOLUME 346

OFEATURE, p 18
NEWS
1 Two in ve NHS professionals think leadership
is poor
New 2.7m genetic testing service for cancer
patients will analyse 97 genes
2 Psychiatrists defend classication of mental illness
diagnoses
Sharing a bed with your baby increases the risk of
sudden infant death
Nicholson cannot shield whistleblowers from action
by foundation trusts
3 Autopsies using MRI in babies and infants are
precise and could help improve rates of consent
4 Choice of Downing Street health adviser prompts
fears about copayments
Integrated care scheme did not reduce emergency
admissions in its rst year
5 Governments plan to raise dementia diagnoses gets
mixed response
Assisted suicide for the dying would reduce
suering, says Falconer
6 Paediatrician wrongly detained in UAE for murder
nally returns home
Cochrane researchers continue to face challenges
over access to data on u drugs
COMMENT
EDITORIALS
7 Premature death among people with mental illness
Graham Thornicroh
O RESEARCH, p 13
8 A call to challenge the Selling of Sickness
Leonore Tiefer, Kim Witczak, and Iona Heath
O PERSONAL VIEW, p 28
9 Suicide among Falkland war veterans
J Holmes et al
10 Preventing admission of older people to hospital
Shaun DSouza and Sunku Guptha
HEAD TO HEAD
16 Should we sequence everyones genome?
As technological prowess soars and costs plummet, is
the era of personalised medicine now in sight? John
Burn says that sequencing everyones genome would
give us unparalleled knowledge to prevent, diagnose,
and treat disease, but there are serious ethical
implications, thinks Frances Flinter
FEATURES
18 DSM-5: a fatal diagnosis?
Unprecedented levels of criticism have marked the
run-up to this weeks launch of DSM-'. Jonathan
Gornall asks whether this mighty US psychiatric
handbook has hnally over-reached itself
ANALYSIS
21 A NICE example? Variation in provision of bariatric
surgery in England
Demand for surgery to treat morbid obesity outstrips
supply. Amanda Owen-Smith and colleagues hnd
regional commissioning policies are not consistent with
NICE guidance and provision of surgery varies widely
RESEARCH
RESEARCH NEWS
11 All you need to read in the other general journals
RESEARCH PAPERS
12 Breast cancer detection and survival among women
with cosmetic breast implants: systematic review
and meta-analysis of observational studies
Eric Lavigne et al
13 The gap in life expectancy from preventable physical
illness in psychiatric patients in Western Australia:
retrospective analysis of population based registers
David Lawrence et al
O EDITORIAL, p 7
14 Completeness and diagnostic validity of recording
acute myocardial infarction events in primary care,
hospital care, disease registry, and national mortality
records: cohort study
Emily Herrett et al
15 Use of serum C reactive protein and procalcitonin
concentrations in addition to symptoms and signs to
predict pneumonia in patients presenting to primary
care with acute cough: diagnostic study
Saskia F van Vugt et al
Media misled over suicides of Falkland combatants, p 9
Access to bariatric surgery varies widely across England, p 21
Study urges against parents sleeping with their babies, p 2
Articles appearing in this print
journal have already been
published on bmj.com, and the
version in print may have been
shortened. bmj.com also contains
material that is supplementary to
articles: this will be indicated in
the text (references are given as
w1, wl, etc) and be labelled as
extra on bmj.com.
Please cite all articles by year,
volume, and elocator (rather than
page number), eg BMJ lu1!;
!/6:flS6.
A note on how to cite each article
appears at the end of each article,
and this is the form the reference
will take in PubMed and other
indexes.
The BMJ is printed on 1uu%
recycled paper (except the cover)
THIS WEEK
Join your
colleagues.
masterclasses.bmj.com
Harry Keen obituary, p 29
COMMENT
LETTERS
24 The science of obesity
25 Research paper of the year award; No doctor
should be untouchable
OBSERVATIONS
FROM THE HEART
26 Dietary advice on added sugar needs
emergency surgery
Aseem Malhotra
MEDICINE AND THE MEDIA
27 Jolie, genes, and the double mastectomy
Richard Hurley
PERSONAL VIEW
28 Admission to hospital
could be considered a
disease
Hugh McIntyre
OBITUARIES
29 Harry Keen
Pioneering diabetes
researcher, staunch NHS supporter, and champion of
people with chronic disease
30 John T D Attenborough; Matthew Walter John Boyd;
Derrick Dexter; Gordon Evison; Alexander Iain Munro
Glen; Nicholas Priestly; Emanuel Tuckman
LAST WORDS
41 Stop protecting the NHS budget Des Spence
A cool response to a hot substance Robin Ferner
EDUCATION
CLINICAL REVIEW
31 Investigation and treatment of imported
malaria in non-endemic countries
Christopher J M Whitty et al
PRACTICE
GUIDELINES
35 Assessment and initial management of feverish
illness in children younger than 5 years:
summary of updated NICE guidance
Ella Fields et al
10MINUTE CONSULTATION
38 Adult acute rhinosinusitis
J Bird et al
ENDGAMES
40 Quiz page for doctors
in training
MINERVA
42 The developing science
of chronotyping,
and other stories
An unusual burn, p 40
Admission a disease? p 28
THIS WEEK
PICTURE OF THE
WEEK
The Arthritis
Research UK Garden
at the 2013 Chelsea
Flower Show. The
garden, designed
by Chris Beardshaw,
is in three sections,
representing the
personal journey of a
patient with arthritis:
from confusion,
represented by the
shady Veiled Garden,
to knowledge,
represented by the
open and formal
Lucid Garden, to
warmth, openness
and confidence
as the patient
learns to manage
the condition,
represented by the
Radiant Garden
(pictured).
R
I
C
H
A
R
D

J
A
N
E

w
w
w
.
r
i
c
h
a
r
d
j
a
n
e
p
h
o
t
o
g
r
a
p
h
y
.
c
o
m
25 May 2013 Vol 346

The Editor, BMJ
BMA House, Tavistock Square,
London WCH JR
Email: editor@bmj.com
Tel: +// (u)lu 7!37 //1u
Fax: +// (u)lu 7!3! 6/13
BMA MEMBERS ENQUIRIES
Email: membership@bma.org.uk
Tel: +// (u)lu 7!3! 69''
BMJ CAREERS ADVERTISING
Email: sales@bmjcareers.com
Tel: +// (u)lu 7!3! 6'!1
DISPLAY ADVERTISING
Email: sales@bmjgroup.com
Tel: +// (u)lu 7!3! 6!36
REPRINTS
UK/Rest of world
Email: ngurneyrandall@bmjgroup.com
Tel: +// (u)lu 3//' '3l'
USA
Email: mfogler@medicalreprints.com
Tel: + 1 (3'6) /39 ///6
SUBSCRIPTIONS
BMA Members
Email: membership@bma.org.uk
Tel: +// (u)lu 7!3! 69''
Non-BMA Members
Email: support@bmjgroup.com
Tel: +// (u)lu 7111 11u'
OTHER RESOURCES
For all other contacts:
resources.bmj.com/bmj/contact-us
For advice to authors:
resources.bmj.com/bmj/authors
To submit an article:
submit.bmj.com
The BMJ is published by BMJ Publishing
Group Ltd, a wholly owned subsidiary of the
British Medical Association.
The BMA grants editorial freedom to the
Editor of the BMJ. The views expressed in
the journal are those of the authors and
may not necessarily comply with BMA
policy. The BMJ follows guidelines on
editorial independence produced by the
World Association of Medical Editors (www.
wame.org/wamestmt.htm#independence)
and the code on good publication practice
produced by the Committee on Publication
Ethics (www.publicationethics.org.uk/
guidelines/).
The BMJ is intended for medical
professionals and is provided without
warranty, express or implied. Statements
in the journal are the responsibility of their
authors and advertisers and not authors
institutions, the BMJ Publishing Group,
or the BMA unless otherwise specified
or determined by law. Acceptance of
advertising does not imply endorsement.
To the fullest extent permitted by law, the
BMJ Publishing Group shall not be liable
for any loss, injury, or damage resulting
from the use of the BMJ or any information
in it whether based on contract, tort, or
otherwise. Readers are advised to verify
any information they choose to rely on.
BMJ Publishing Group Ltd lu1!
All Rights Reserved. No part of this
publication may be reproduced, stored in
a retrieval system, or transmitted in any
form or by any other means, electronic,
mechanical, photocopying, recording, or
otherwise, without prior permission, in
writing, of the BMJ.
Published weekly. US periodicals class
postage paid at Rahway, NJ. Postmaster:
send address changes to BMJ, c/o Mercury
Airfreight International Ltd Inc, !6' Blair
Road, Avenel, NJ u7uu1, USA. 5796.
Weekly.
Printed by Polestar Limited
BMJ. COM POLL
Last weeks poll asked:
Should governments introduce plain packets
for cigarettes?
72% voted yes
(total 1002 votes cast)
BMJ lu1!;!/6:f!u69
This weeks poll asks:
Should we sequence everyones genome?
Head to Head:
Yes BMJ lu1!;!/6:f!1!!
No BMJ lu1!;!/6:f!1!l
Vote now on bmj.com
MOST SHARED
Helping patients to die well
Seeing double: the low carb diet
Are MOOCs the future of medical education?
Orlistat: should we worry about liver
inflammation?
Measles and stroke show why healthcare must
innovate
RESPONSE OF THE WEEK
It is extraordinary how little there is written
about the emotional labour of care in relation
to doctors. Menzies Lyths famous work looked
at the social defences used to protect nurses
from the burden of emotional involvement
separating care into parcels for different nurses
to deal with medications, washing, observations,
etc. In the same year (1959) Balint used the term
collusion of anonymity to describe the manner
in which doctors could claim responsibility for
their organ of interest, but none would take
responsibility for the patient.
Increasing specialisation, working time
directives, financial contracts, and disease
pathways all act as social defences. Combine this
with the head-in-the-sand approach that typifies
the medical professions attitude to its own
subjectivity and involvement with care and you
can see why compassion is so endangered.
Jonathon P Tomlinson, GP, London, UK, in response
to How to encourage compassion
(BMJ 2013;346:f2049)
THIS WEEK
Is that really the right title? asked a colleague at a
recent editorial meeting, spotting this weeks Personal
View on the list of forthcoming articles. See what you
think. In Admission to hospital could be considered a
disease, Hugh McIntyre argues that being in hospital
doesnt simply put patients at risk of further illness
and early readmission; rather, it is itself an abnormal
health status associated with physiological and
psychological stresses (p 28). This is echoed in the
editorial by Shaun DSouza and Sunku Guptha (p 8).
They dene frailty as a dynamic condition associated
with sudden profound decompensation secondary to
a stressor, and nd no good evidence that admission
or readmission to hospital is reduced by changes
in community care for frail older people. McIntyre
suggests that we label hospitalisation as a disease,
monitor and study it more closely, and focus on its
prevention and treatment.
McIntyres concept seems analogous to
institutionalisation. Is that syndrome listed in the
Diagnostic and Statistical Manual of Mental Disorders
(DSM), I wonder? If I had $199 (131, t155) to spare
I could look it up in the latest edition, DSM-5, which
was published last week. Jonathan Gornall reviews
the background to this highly controversial tome and
interviews its main detractors and defenders (p 18).
Shitij Kapur, dean of the Institute of Psychiatry at Kings
College London, says that critics can nd all the faults
with DSM but are totally naive about the total mess
that the diagnosis of psychiatric disorders used to be
about /u years ago . . . What you called schizophrenia
in New York wasnt schizophrenia in London, and we
have made a huge transition since then. The next
transition, hopes Thomas Insel, director of the US
National Institute of Mental Health, will reflect the
biological basis of mental disorders much more closely.
But we may be onto DSM-1u by the time the institutes
biomarker focused, Research Domain Criteria project
bears fruit.
For good medical practice, we need tests that
answer a specic question with a reasonable prospect
of useful results, counsels Frances Flinter in this
weeks Head to Head debate (p 16). Its hard to argue
against that. She also argues that possessing the
technical ability to do something new doesnt justify
going ahead with it, especially in an ethically complex
area like whole genome sequencing in healthy
individuals. Flinter believes that it has nothing to oner
clinically at the moment because most of the data
it generates are meaningless. John Burn disagrees.
We can have a whole genome for the price of a family
package holiday, he explains, and theres scope right
now for cheap, fast genotype panels in some frontline
care settings. But are we ready for a mixed economy
of stored whole genomesand the labels that will
come with it? Not if this means dumping heaps of
molecular uncertainty on patients, families, and their
carers, concedes Burn.
Trish Groves, deputy editor, BMJ
tgroves@bmj.com
Cite this as: BMJ 2013;346:f3352
OFollow Trish Groves at twitter.com/trished
EDITORS CHOICE
Too many labels
Being in hospital
doesnt simply put
patients at risk of
further illness and
early readmission;
rather, it is itself an
abnormal health
status associated
with physiological
and psychological
stresses
Sign up today using
your smartphone
follow these steps:
ODownload a free QR reader
from your handset's app store
OHold your smartphone over
the QR code
OYou will then be forwarded
to the email sign up page
Podcasts
Listen to what you want,
when you want, and how you want
Imagine waking up and automatically having the latest
medical news, views, research, and education to listen to
for free while you commute to work or exercise. This is the
podcast listening experience.
Sign up at iTunes or download from
bmj.com/multimedia/podcasts
OTo receive Editors Choice by email each week,
visit www.bmj.com/newaccount
NEWS
BMJ | 25 MAY 2013 | VOLUME 346 1
Matthew Limb LONDON
Many senior health professionals and managers
doubt the ability of NHS leaders to improve the
quality of patient care, says the health think tank
the Kings Fund.
In a survey of some 900 NHS professionals in
England carried out aher the Francis inquiry into
Mid Staordshire NHS Foundation Trust, almost
three quarters (73%) said that they didnt think
that quality of care was given enough priority.
Only 14% thought that the quality of leader-
ship in the NHS as a whole was good or very
good, while 40% thought it poor or very
poor.
The Kings Fund said that NHS boards should
be doing much more to exercise clear and visible
leadership to improve the quality of care their
organisations provide. It also urged NHS leaders
in charge of the new fragmented national bod-
ies to set a consistent tone and clear standards
and goals to improve quality and patient safety.
Unless these organisations demonstrate an
uncommon commitment and ability to work
together, there is a real risk of NHS sta receiv-
ing mixed messages, with a lack of clarity and
direction on what really matters, it said. In this
scenario, a repetition of the failures that occurred
at Mid Staordshire is more likely.
Nicola Hartley, director of leadership devel-
opment at the Kings Fund, said that it was the
responsibility of all NHS organisations and pro-
fessionals to put the needs of patients above
Two in five NHS professionals think leadership is poor
Krishna Chinthapalli BMJ
A new programme to develop genetic
testing in patients with cancer will be
piloted in London within the next six
months.
The Mainstreaming Cancer
Genetics programme has been
launched with 2.7m (3.2m;
$4.1m) funding from the Wellcome
Trust and will use next generation
gene sequencing technology to
test patients with cancer. The test,
called the TruSight cancer targeted
microarray panel, costs a few
hundred pounds per patient and
will analyse 97 cancer predisposition
genes for mutations.
Currently no test is available
for over half these genes, and the
microarray panel could also replace
testing for mutations in single genes,
such as BRCA1. It was a BRCA1
gene mutation that led the US actor
Angelina Jolie to recently undergo
double mastectomy to try to prevent
future breast cancer.
The new programme was
announced at a press briefing on 20
May by a team from the Institute of
Cancer Research, London, and the
Wellcome Trust.
Nazneen Rahman, lead
investigator of the programme and
head of genetics at the Institute of
Cancer Research, emphasised that
the programme was not related to
the publicity about Jolies decision.
However, he added, It is very
important to know if a mutation in
a persons genetic blueprint has
caused their cancer. It allows more
personalised treatment . . . For
example, such people are often at risk
of getting another cancer and may
choose to have more comprehensive
surgery or different medicines.
By 2014 the programme will pilot
routine microarray panel testing
in women with breast and ovarian
cancers at the Royal Marsden NHS
Foundation Trust. Martin Gore, the
trusts medical director, said, This
programme will help make genetic
testing quicker and simpler, and the
Royal Marsden is excited.
Cite this as: BMJ ;:f
those of the organisation, team, or profession.
Our survey suggests that we have a long road
to travel to achieve this, she said.
The survey hndings appear in a report, Patient-
Centred Leadership: Rediscovering our Purpose,
published by the Kings Fund on 23 May.
1
The
survey was carried out in February and March
this year aher the inquiry by Robert Francis QC
condemned shocking failures of care at Mid
S taordshire.
Professionals were asked about the factors
aecting the quality of patient care and the role
of leadership in delivering improvements.
Almost two thirds (62%) of the 900 respond-
ents worked in an acute setting. Nurses were the
largest professional group to respond (36%), fol-
lowed by managers (20%) and directors or heads
of service (17%).
More than a third of respondents (36%) said
the quality of leadership in their organisation
was good or very good, while 26% said it was
poor or very poor. Eleven per cent thought
the quality of leadership in their own service
or team poor or very poor. When asked to
name the biggest barrier to increasing the qual-
ity of care, 40% of NHS professionals cited time
and/or resources.
Organisational culture was next (28%),
though this was identihed as the most impor-
tant factor by respondents who were executive
directors (48%).
The Kings Fund said that although the
respondents clearly saw a role for clinicians and
managers in improving the quality of care, they
were less clear on the role of boards.
New 2.7m genetic testing service for cancer patients will analyse 97 genes
Nicola Hartley said patients needs must come first as outlined in the report from Robert Francis (right)
UK news Choice of Downing Street adviser prompts fears that copayments may be on the agenda, p
World news Paediatrician wrongly detained in UAE for murder finally returns home, p
References on news stories are in the versions on bmj.com
bmj.com
Surgeon has
conditions
placed on his
practice
NEWS
Nigel Hawkes LONDON
Advice to mothers that it is safe to sleep with their
babies should be reconsidered, say the authors of a
new study, which shows that it multiplies the risk of
sudden infant death.
1
Current recommendations in the United Kingdom
are that it is safe to share a bed with a baby so long
as neither parent smokes or has taken alcohol or
drugs. But a new analysis of data from five previous
studies, published in BMJ Open, has found that
even if these conditions apply, the risk of the baby
dying is much higher when bed sharingfive times
higher for babies under 3 months old.
The authors, led by Robert Carpenter of the
London School of Hygiene and Tropical Medicine,
estimate that 88% of sudden infant deaths that
occur while bed sharing would not have occurred
Sharing a bed with your
baby increases the risk of
sudden infant death
2 BMJ | 25 MAY 2013 | VOLUME 346
Zosia Kmietowicz BMJ
Specialists in both psychiatry and psychology
came to the defence of classihcation systems for
mental illnesses last week, aher clinical psycholo-
gists called for diagnostic labels to be abandoned
in favour of a conceptual system that was not
based on a disease model.
The British Psychological Societys Division
of Clinical Psychology, which represents more
than 10 000 practitioners, issued a statement
on 17 May damning diagnoses such as schizo-
phrenia, personality disorder, and bipolar dis-
order because they have limited reliability and
q uestionable validity.
1
It added: Psychiatric diagnosis is ohen pre-
sented as an objective statement of fact, but is,
in essence, a clinical judgement based on obser-
vation and interpretation of behaviour and self-
report, and thus subject to variation and bias.
Instead, the division said that there needed to
be a paradigm shih in relation to functional psy-
chiatric diagnoses . . . [with] an approach that is
multi-factorial, contextualises distress and behav-
iour, and acknowledges the complexity of the
interactions involved in all human experience.
But Nick Craddock, professor of psychiatry at
the University of Cardi, said that while no clas-
sihcation system was perfect, diagnostic rigour
underpinned both mental health research and
clinical practice.
He spoke on 17 May at a journalists brief-
ing on psychiatric classihcations, ahead of the
publication of the hhh edition of the Diagnostic
and Statistical Manual (DSM-5) at the American
P sychiatric Associations annual
meeting later that day.
2
I have been a
major critic of the
DSM approach. But
al t hough t her e
a r e s h o r t -
comi ngs, I
still believe
absolutely that we must maintain the scientihc
and thoughtful approach where we try to bring
together evidencepsychological, social, bio-
logicaland deliver the best help we can. This is
absolutely the wrong time to be going back to the
40 year old argument [of psychiatric classihca-
tion]. It is really absolutely ridiculous, he said.
Craddock said he did not believe that science
had advanced sumciently in the 17 years since
the DSMs fourth edition was published in 1994
to warrant the new edition.
The hxation in the media on quirky new diag-
noses in DSM-5, such as internet addiction,
shyness in children, and hypersexual disorder,
is not helpful. These are not that important to
clinicians. What is important to them are subtle
changes, such as a new criteria for diagnosing a
core disorder, said Craddock.
David Clark, professor of experimental psy-
chology at the University of Oxford and a fellow
of the British Psychological Society, said that the
statement from the clinical psychologists com-
pletely ignored the fact that the mental health
classihcation systems relied equally on underly-
ing psychological processes and biological ones.
He said that distinctive diagnoses had led to
targeted research and new treatment approaches.
In the late 1980s, anxiety was simply classihed
as being phobic neurosis or anxiety neurosis. But
now there is also panic disorder, post-traumatic
stress disorder, social anxiety disorder, and
obsessive compulsive disorder, said Clark.
If we abandoned classihcation, it would be
dimcult to know what treatment to give, he told
the briehng.
[Post-traumatic stress disorder] is about intru-
sive memories, and to get good results you have
to use very specialised memory related therapy
techniques, but those techniques have no rele-
vance in social anxiety disorder, where video
feedback is much more powerful.
OFEATURE, p 18
Psychiatrists defend classification
of mental illness diagnoses
Catherine Zeta-Jones (left), Alastair Campbell, and Ruby
Wax have been vocal about their mental health issues.
But psychologists argue that diagnoses are unreliable
Nicholson cannot shield
whistleblowers from
action by foundation trusts
Clare Dyer BMJ
Doubts have been raised about the value of NHS
chief executive David Nicholsons pledge to MPs
to intervene with NHS organisations if he caught
a whi of failure to support whistleblowers. In
a letter seen by Channel 4 News, he has admitted
that his powers to step in are limited.
When asked about NHS trusts not provid-
ing support to whistleblowers, Nicholson told
the Commons Health Committee in February,
Wherever I see it or if I have a whi of it, I imme-
diately intervene.
But in a letter just days later to a whistleblow-
ing doctor with a foundation trust who signed a
gagging agreement aher raising concerns about
treatment of patients on a dementia ward,
Nicholson told her that he could not step in. This
was because a legal process had concluded,
and because foundation trusts are separate legal
bodies from the Department of Health.
The agreement included the usual clause
spelling out the employees right under the
Public Interest Disclosure Act (PIDA) to raise
concerns about patient safety and care, but said
that this applied in respect only of matters com-
ing to the employees attention aher the date of
this agreement. In fact, such a limitation would
probably be void under the act, but Channel 4
said that the doctor was unwilling to test it.
NEWS
if bed sharing had been avoided. They say that
over the past 10 years, there has been a marked
increase in bed sharing, prompted by charities
such as the National Childbirth Trust. According to
the trust, bed sharing aids breast feeding and the
risks are probably negligible if both parents avoid
smoking, alcohol, and drugs, and if they sleep in a
bed rather than on a sofa.
Earlier studies calculated the risks of sudden
infant death syndrome differently, which made the
risks difficult to compare. As a result, the authors of
the new study went back to the raw data from five
case-control studies to create a combined dataset
of 1472 cases of sudden infant death syndrome
and 4679 controls. The data came from a trans-
European study and national studies in Scotland,
New Zealand, Ireland, and Germany, and included
babies aged up to 1 year.
The authors concluded that even in the best of
circumstances (that is, breastfed infants and no
parental risk factors), the risk of sudden infant death
syndrome was 2.7 times higher (95% confidence
interval 1.4 to 5.3) if babies shared a bed with
parents rather than sharing a room. In the worst
circumstances (that is, bottle fed babies and both
parents were smokers and drinkers), this risk was
15.6 times higher (5.7 to 41.5).
The current recommendations are not effective
and should be modified to take a more definitive
stance against bed sharing for babies under 3
months, the authors wrote. We do not suggest that
babies should not be brought into the parents bed
for comfort and feeding. This has been investigated
in previous studies and has not been found to be a
risk factor, provided the infant is returned to his or
her own cot for sleep.
Ullat aut ute voloremquid et faciundi quat.
Doluptat delis sit, volores et et
BMJ | 25 MAY 2013 | VOLUME 346 3
Recommendations should take a stronger stance against bed sharing for babies under 3 months
A
L
E
X

L
I
N
G
H
O
R
N
/
G
E
T
T
Y

I
M
A
G
E
S
A large study has found that
postmortem examinations in
fetuses and infants up to 12
months using a combination
of magnetic resonance imag-
ing (MRI) and blood tests are
as accurate as conventional
methods.
The authors believe that
the results provide sumcient
evidence for minimally invasive autopsies to be
oered routinely to parents whose babies have
died, and that this could help to improve autopsy
rates that have been falling for the past 40 years.
The use of postmortem examinations using
MRI scans is currently restricted to only a hand-
ful of centres worldwide. Small scale studies have
been reporting for more than a decade, but the
accuracy of the method is uncertain.
Researchers from University College London
and Great Ormond Street Hospital therefore ana-
lysed the results from 400 fetuses, babies, and
children who were undergoing conventional post-
mortem examinations, and compared the cause
of death with that obtained from whole body,
postmortem MRI alone, or with other minimally
invasive methods. These methods included blood
samples taken by needle, visual examination of
the body, and genetic and metabolic tests, but did
not include tissue biopsies.
The scans took place aher 6
pm and before 8 am, so did
not aect the access to MRI.
Reporting in the Lancet, the
researchers found that autop-
sies using MRI alone resulted
in the same cause of death as
conventional autopsies in 222
of 400 cases (55.5% (95%
conhdence interval 50.6% to
60.3%)). But when MRI was combined with the
other tests, the results concurred in 357 cases
(89.3% (85.8% to 91.9%)).
1
They also found that concordance varied with
age, with the highest rate for fetuses (95-96%),
newborn babies up to one month (81%), and
infants up to 12 months of age (85%). For chil-
dren aged between 12 months and 16 years, con-
cordance was just 54%.
The dierences were largely because deaths in
fetuses and infants tended to be a result of struc-
tural abnormalities, while older children died
from undetected pneumonia, myocarditis, and
sepsis, said the researchers.
Sudhin Thayyil, consultant neonatologist at
Great Ormond Street Hospital and University
College Hospital and lead author, said that in
40% of cases an MRI scan was enough to decide
to an accuracy of 99.4% whether a conventional
postmortem examination would deliver anything
interesting. In other cases, keyhole surgery could
be used to retrieve samples of tissue.
He also said that autopsies using MRI could
negate the need for microscopic brain examina-
tion when brain MRI is normal.
The Department of Health, which funded
the study, has commissioned a report of how to
establish a national service for minimally inva-
sive postmortem examinations. Andrew Taylor,
consultant radiologist at Great Ormond Street and
a study author, believes that there could be six
expert centres in England to which local pathol-
ogy and radiology services could feed MRI scans.
Between 2000 and 2007, consent rates in the
United Kingdom for fetal autopsy fell from 55% to
45%. The rate for neonatal autopsy fell from 28%
to 21%, despite more parents being oered them,
largely as a result of the retained organ scandal
at Alder Hey hospital that was reported in 1999.
Taylor told the briehng: If we oer something
less invasive, [parents] are more willing to take it
up. Minimally invasive autopsy also has potential
for religious groups that do not agree with con-
ventional autopsy.
He added: Autopsies not only help us to
establish the cause of death, but they ohen play
an important role in advancing medical research
and knowledge.
Autopsies using MRI in babies and infants are
precise and could help improve rates of consent
Retained organs at Alder Hey hospital
in 1999 led to a drop in autopsies
NEWS
Ingrid Torjesen LONDON
Fears that the appointment of Nick Seddon as
the new special adviser on health at Downing
Street is a sign that the government is consid-
ering the introduction of copayments for NHS
patients have been dismissed by senior health
policy hgures who know him well.
Before being invited to advise on health
and social care as part of a new team led by Jo
Johnsonthe brother of London mayor Boris
JohnsonSeddon was deputy director of the
right wing think tank Reform, which promotes
free market policies and small government.

4 BMJ | 25 MAY 2013 | VOLUME 346
Nick Seddon has said that some hospitals need to
close, but will he succeed where others have failed?
Civil servants to spend a month on
NHS frontline: Everyone working in the
Department of Health for England will have to
spend time every year at the NHS frontline to
see how policies work in practice, the health
secretary, Jeremy Hunt, has said. Directors
and senior civil servants will spend at least
a month each year learning about patients
experiences. The move is a response
to the Francis report into failings at Mid
Stanordshire NHS Foundation Trust.
Gay men in New York urged to have
meningitis vaccine: Medical omcials in
New York City have recommended that
men who have sexual contact with other
men are vaccinated against meningococcal
meningitis, afer a cluster of cases. Twenty
two cases of Neisseria meningitidis strain C,
including seven deaths, have been reported
among men who have sex with men since
2u1u. Four of the cases occurred in 2u13.
Kings Fund calls for joint NHS and social
care budget: The governments proposals for
social care in the Care Bill will not solve the
problem of funding of social care in England,
a report from the think tank the Kings Fund
has said.
1
The bill introduces a cap on the
amount that individuals contribute towards
the cost of their care, but the report says that
more people need to be eligible for help. It
calls for the 2u15-16 spending review to be
used to deliver a single budget for the NHS
and adult social care.
Doctors need to pass Hebrew tests to work
in Israel: Doctors who wish to migrate to
Israel will have to take tests to prove they can
converse with patients in basic Hebrew, the
Knessets Labour, Social Anairs and Health
Committee has decided. Nurses must already
prove their prociency in Hebrew, but a
number of malpractice lawsuits showed that
doctors too need to be able to be understood
in the countrys primary language.
Rise in incidence of dengue
fever prompts warning:
Public Health England
is urging travellers to
protect themselves
against mosquito bites
afer gures showed 1/1
cases of dengue fever in
travellers returning to England,
Wales, and Northern Ireland between January
and April in 1u13, nearly triple the 51 in the
same period in 2u12.
IN BRIEF
Choice of Downing Street health adviser
prompts fears about copayments
Before joining Reform, Seddononly in
his mid-30swas head of communications at
C ircle, the hrst private company to win a fran-
chise to run an NHS hospital, Hinchingbrooke
in C ambridgeshire.
Seddon has been reported to have argued
that, based on international evidence, imposing
charges for services such as seeing a general prac-
titioner could be eective.
1
He said this in 2012,
when asked to comment by the BBC on a report
from the Institute for Fiscal Studies. The report
had concluded that the government needed to
consider more radical steps because of rising
demands on the NHS from the ageing population,
such as introducing charging for some services,
stopping low priority care, or raising taxes.
2
Nick Bosanquet, professor of health policy
at Imperial College London, and a member of
Reforms advisory board, said that fears about
copayments were quite misplaced. [Seddon]
has said no more than Malcolm Grant [chair] of
NHS England
3
that at some point we may have
to consider user chargeshe didnt say that that
would be desirable or that it is going to hap-
pen soon, but realistically it may happen in the
future, he told the BMJ.
Gareth Iacobucci BMJ
An evaluation of the hrst year of a major inte-
grated care pilotseen as a test bed for national
government policyhas found little change in
emergency admissions and in the wider use of
health and care services.
But researchers from the Numeld Trust and
Imperial College London
1
said that more time
was required to assess the longer term eect
of the integrated care pilot in inner northwest
L ondon, which has focused on producing indi-
vidual care plans for people with diabetes and
those older than 75.
The report follows the launch of the govern-
ments new flagship policy to fully integrate
health and social care in England within hve
years,
2
and its results will be keenly monitored
by commissioners and policy makers looking to
implement new projects across England.
The evaluation said that the pilot found good
progress in designing and implementing a com-
plex intervention, but the hrst group of 1236
Integrated care scheme
did not reduce emergency
admissions in its first year
patients treated under the new case manage-
ment arrangements had not shown any sig-
nihcant reduction in emergency admissions or
signihcant changes in the wider use of services.
It said that cases would need to be followed
for longer to assess whether the pilots were
aecting emergency hospital admissions and
total costs.
The pilot uses a new information technology
tool that allows patients to be stratihed accord-
ing to risk; shares plans across dierent care
settings; identihes those needing intensive case
management; and enables the implementation
of plans to be monitoredusing data from gen-
eral practices, acute trusts, mental and commu-
nity healthcare providers, and social care.
3
Regular multidisciplinary meetings review
and plan peoples care, aided by a newly formed
integrated management board of sta, patient
representatives and providers, and project man-
agement and support from a dedicated team.
Researchers reported a high level of commit-
ment to the pilot among healthcare profession-
als, in particular to the care planning process,
which they said had led to improved collabora-
tion across dierent parts of the local health and
social care system.
NEWS

BMJ | 25 MAY 2013 | VOLUME 346 5
M
A
R
K

T
H
O
M
A
S
Charles Falconer (left) introduced the bill to legalise assisted suicide, but it would not have helped the
late Tony Nicklinson (centre), nor will it help Paul Lamb, who is paralysed but not terminally ill
Assisted suicide for the
dying would reduce
suffering, says Falconer
Clare Dyer BMJ
A private members bill to legalise physician
assisted suicide for terminally ill patients in Eng-
land and Wales has been tabled in the House
of Lords by a former Labour lord chancellor,
Charles Falconer.
The bill, which had its formal rst reading on
May, follows the report last year of the inde-
pendent Commission on Assisted Dying, chaired
by Falconer, which branded the existing law
inadequate and incoherent.
The bill was introduced in parliament in the

UK is the same week that lawyers for two para-
lysed men argued in the Court of Appeal that
doctors should be allowed to help them end their
lives. The argument was rejected last year in the
High Court, where two judges ruled that it was a
matter for parliament, not the courts.
The bill would not have helped the late Tony
Nicklinson, who refused food and died after
unsuccessfully challenging the law, or Paul
Lamb, one of the two men currently seeking the
right to die. Neither man was terminally ill, and
both were too paralysed to assist in their own
suicide. A doctor would have had to actively end
their lives, which would amount to murder.
The bill sets up a framework under which two
doctors would have to certify that a patient was
likely to live for six months or less, was mentally
competent, and had a clear and settled intention
to die. The patient would be given a lethal dose
of drugs to self administer.
A private members bill is highly unlikely to
become law without the support of the govern-
ment, which regards the matter as an issue of
conscience for each MP. However, health minis-
ter Anna Soubry told the Times newspaper last
year: I think its ridiculous and appalling that
people have to go abroad to end their life instead
of being able to end their life at home.
Fellow health minister Norman Lamb said
at the time that there was a case for review of
the law, while indicating that it was a matter
for parliament, not the government. The bill is
expected to have its rst detailed debate in the
House of Lords in the autumn.
A poll of people in the UK, published
this week by Cicero Research, found that %
supported the legalisation of assisted dying, but
only % of those surveyed believed the right
should be restricted to terminally ill people.
Polls show that most doctors oppose the move,
as does the BMA.
Seddon is an expert on innovation in health
services and would bring some fresh and badly
needed ideas for lowering costs, Bosanquet
added. Hes very keen on improving communi-
cation with patients and making it much easier
for patients to get direct access to their [general
practitioners] by telephone, email, and texting,
he said.
In an article for the Health Service Journal last
year,
Seddon wrote: If we want the NHS to

survive we will need to transform the relation-
ship between the individual citizen and the state
with personal technology aiding the exercise of
personal power, control and accountability. He
added, Services will need to be shaped around
patients in a complete overhaul of the business
model, which included a reduced emphasis on
hospital care, more care in community settings,
use of new technology, personal budgets, and
helping services do more with less.
Part of this overhaul required workforce
reforms, with local rather than national terms
and conditions, and more skill mixed, team
based working.
Seddon has acknowledged that some hos-

pitals need to close, but, as Bosanquet points
out, he is not the rst person to say this. The
question is whether Seddon can drive the gov-
ernment into seeing this through where others
have failed.
Governments plan to raise number of
dementia diagnoses gets mixed response
Susan Mayor LONDON
NHS England has set its first
national target to improve the
rate of diagnosis of dementia,
with the aim of ensuring that by
2015 two thirds of people with
the condition have a diagnosis.
Its latest report, published on 15
May, also says that appropriate
post-diagnosis support should
be available.
1
The diagnosis target will mean
increasing the rate of diagnosis
from the current national average
of around 45%, which the health
secretary said was shockingly
low, to match the 67% achieved
in the best performing local area.
The result will be an additional
160 000 diagnoses each year
over the 2011-12 figure.
The report outlines a new
enhanced service that GPs can
provide as part of the GP contract
for 2013-14. This rewards
practices for having a proactive,
case finding approach to the
assessment of patients who
may be showing early signs of
dementia.
But Peter Gordon, a consultant
psychiatrist for older adults in
Sauchie, Clackmannanshire,
said that making a diagnosis of
dementia had been presented as
if it had only benefits. As with any
intervention, he said, there were
also potential harms that should
not be ignored.
These included that:
No more than half of people
with mild cognitive loss
will develop dementia. The
risk of misdiagnosis in this
group is high and may result
in significant harm and the
diversion of resources away
from the people most in need
Targets in this area too easily
ignore the right of informed
consent and the ethical
principle of doing no harm
The unintentional message
given to society is that all
memory loss in older adults
is early dementia or early
Alzheimers disease.
The Royal College of General
Practitioners said that the
governments call for a rise in
diagnosis must be backed by
appropriate funding and that
patients wishes should be
paramount.
The colleges chairwoman,
Clare Gerada, said, Not every
person with dementia will
find that the advantages of an
early diagnosis outweigh the
possible disadvantages, and
evidence shows that if a persons
wellbeing is not enhanced by
receiving a diagnosis then it
should not be forced on them.
NEWS
Cochrane researchers continue to face
challenges over access to data on flu drugs
Cochrane Collaboration researchers have said
that they may be unable to analyse clinical trial
data given to them by GlaxoSmithKline (GSK)
on the antiviral drug zanamivir (marketed as
Relenza) because nearly two thirds of the reports
provided have been redacted in dierent ways.
The researchers have been trying to get access
to data held by drug companies on neuramini-
dase inhibitors for more than three years so that
they can review the drugs safety and emcacy.
GSK was the hrst company to hand over clini-
cal study reports, giving researchers 30 reports at
the end of April. The reports hold vast amounts
of information on trials conduct and outcomes.
The Cochrane researchers remain positive
about GSKs attempt to share the detailed data it
holds on zanamivir but have written to the com-
pany asking it to clarify why certain information
had been redacted.
1
This information includes
descriptions of serious adverse events and the
dates of the start and end of studies.
The letter to GSK from the Cochrane investiga-
tors Tom Jeerson and Peter Doshi said, If we do
not have a complete CSR [clinical study report],
we have to hgure out what we are missing and
why. This is not something we like doing, as we
are neither auditors nor regulators.
Jefferson and Doshi asked 19 questions of
GSK about the data and the purpose of releasing
them to researchers, including, What type of
independent research does GSK think this kind
of release enables?
The researchers are trying to dehne a mini-
mum dataset for use in evidence synthesis but
acknowledge that the learning curve for them
and GSK is steep. They hope that a discussion
will lead to a more fruitful analysis of the data.
In February GSK became the hrst drug com-
pany to sign up to the AllTrials initiative (alltri-
als.net), set up in January by the charity Sense
About Science, the BMJ, and other supporters
of transparency in research.
2
It is calling for the
registration of all clinical trials and for full study
results and full clinical study reports to be made
publicly available.
But the cat and mouse game of hiding data and
promising disclosure continues.
On 2 April the drug company Roche told the
Cochrane researchers that it would give them 74
clinical study reports on oseltamivir (Tamiu)
over the next few months.
3
But the researchers
have not yet received any of the studies.
Jeerson told the BMJ, We have already had
one promise from Roche in December 2009 in the
BMJ. We are still waiting for the full study reports
that were promised. So far there have been prom-
ises, declarations, and statements, but we have
not seen anything.
On 10 May Jeerson wrote to Roches life cycle
leader on oseltamivir aher analysing the clini-
cal study reports on two trials that the company
sponsored.
4
He called for one 14 year old study
to be published for the whole world to beneht
from the data contained in it. The second study,
which was published in JAMA in 2001,
5
needed
to be retracted or amended because many of the
statements and conclusions it drew were not
supported by the design of the trial or the data
presented, he said.
That study has played and continues to play a
role in the clinical recommendations for the use
and the stockpiling of Tamiu at huge taxpayers
expense, he wrote.
Campaigners for full transparency of clinical
data are also frustrated by an interim ruling by
the General Court of the European Union order-
ing the European Medicines Agency not to release
documents it holds on the safety and emcacy of
drugs made by the US companies AbbVie and
InterMune.
Sophie Arie LONDON
Cyril Karabus, the 78 year old South African doc-
tor detained in the United Arab Emirates (UAE)
for allegedly murdering a patient there, has
hnally returned home aher a nine month ordeal.
The paediatric oncologist was arrested in
August 2012 while in transit in Dubai and was
told that he had been convicted in absentia for
killing a patient in Abu Dhabi while working as
a locum a decade earlier.
He denied all charges and, aher months of
delays, a court acquitted him of manslaughter
and forgery of patient records in March. The
professor was initially jailed for 57 days and
appeared for several court hearings in shack-
les. But the prosecution was unable to present
any evidence that he had caused the death of
a young Yemeni patient with leukaemia whom
he had treated.
Karabus, who has a heart con-
dition but managed to maintain
relatively good spirits through-
out, believes that his conviction
was concocted as part of a process
for awarding blood money to the
family of the deceased child. He
has told South African media that
court documents on the original
2002 ruling were full of lies,
including a statement that he had
been informed of the verdict.
Karabuss case has highlighted
the risks for foreigners of working in the UAE,
who are attracted by high salaries and are
thought to comprise more than 90% of the
states medical workforce. Several other medi-
cal professionals are known to have received
similar convictions in absentia and have lim-
ited their travel to avoid arrest.
The World Medical Association,
which had campaigned for Kara-
buss release, warned that it would
continue to advise doctors about
the risks of working in the UAE.
Before his ordeal, the retired pro-
fessor, who is widely respected in
South Africa, had travelled exten-
sively on locums to top up his small
South African pension. He has now
incurred around E70 000 (t83 000,
$107 000) in legal fees alone. His
lawyer, Michael Bagraim, told local media that
he would be looking for some new locum work
soon.
I dont believe hell ever go back to the UAE,
but he might locum in England, he said.
6 BMJ | 25 MAY 2013 | VOLUME 346
F
R
A
N
C
I
S
C
O

K
J
O
L
S
E
T
H
/
A
P
/
P
A
Paediatrician wrongly detained in UAE for murder finally returns home
Cyril Karabus was jailed
for 57 days and detained
in UAE for nine months
The researchers have written to GSK to ask why
certain information on Relenza was redacted
BMJ | 25 MAY 2013 | VOLUME 346 7

Editorials are usually commissioned. We are, however, happy to consider and peer review unsolicited editorials
See http://resources.bmj.com/bmj/authors/types-of-article/editorials for more details
EDITORIALS
Premature death among people with mental illness
At best a failure to act on evidence; at worst a form of lethal discrimination
Graham Thornicroft professor of community psychiatry,
Health Service and Population Research Department, Kings
College London, Institute of Psychiatry, London SE' 3AF, UK
graham.thornicroft@kcl.ac.uk
The hndings of a linked paper by Lawrence and
colleagues raise disturbing questions about our
disregard for the duration and value of the lives
of people with mental illness.
1
It has been clear
for more than 50 years that people with the more
disabling forms of mental illness do not live as long
as those without mental illness.
2
This hnding has
been repeatedly reported across decades and con-
tinents.
3

4
Evidence from low and middle income
countries is sparse,
5

6
despite poignant accounts
from colleagues in such countries of people with
severe mental illness being abandoned in forests
or deserts when families can no longer cope. Law-
rence and colleagues research covers new ground
by focusing on reduced life expectancy among
p eople with a broader range of mental illnesses,
rather than those with severe conditions only.
However, their results support the accumulating
evidence that inequalities between those with and
without mental illness are not improving.
For the past 20 years, the mortality gap among
people with mental illnesses (around 15 years for
women and 20 years for men) has not been closing
in Australia or Scandinaviatwo of the most amu-
ent parts of the world with among the most accessi-
ble health systems.
1

7
We are coming to understand
that this excess mortality is not the result of higher
suicide rates, but rather a combination of socioeco-
nomic, healthcare, and clinical risk factors.
8
What are the implications of these sobering hnd-
ings? Three things are clear. Firstly, the time for
descriptive research alone is over. We now need evi-
dence based interventions that can reduce excess
mortality. Such interventions may include smok-
ing reduction or cessation in people with mental
illness, or specihc lifestyle programmes that seek to
modify risk factors for cardiovascular, respiratory,
and malignant diseases.
9
Secondly, because of the sheer scale of the prob-
lem, it needs to be placed high on the public health
priority list. Here there are signs of progress. In the
United Kingdom, the Health and Social Care Act
2012 included a commitment to parity of esteem
between mental and physical health, and as a con-
sequence the NHS Mandate of 2012 requires the
NHS to tackle disparities between mental and
physical healthcare. Internationally, the World
Health Organization Mental Health Action Plan,
to be presented to the World Health Assembly in
late May 2013, states as its overall goal: To pro-
mote mental wellbeing, prevent mental disorders,
and reduce the mortality and disability for persons
with mental disorders.
10
Another vital opportunity
arises as David Cameron (UK), Ellen Johnson Sirleaf
(Liberia), and Susilo Bambang Yudhoyono (Indone-
sia) co-chair the high level panel on the post 2015
development agenda, to set the targets to follow the
millennium development goals. This panel must
fully include the interests of people with mental
illness within its global vision.
Thirdly, this huge loss of life among people with
mental illness needs to be recognised as a human
rights disgrace. Despite the right to health having
become law in 126 countries worldwide (through
their ratihcation of the United Nations Convention
on the Rights of Persons with Disabilities),
11
it has
not been implemented into eective action. Meas-
urable progress has now been made in treatment
coverage for people with major global communi-
cable diseases, such as HIV. However, there are no
internationally agreed standards to measure how
many people with mental illnesses need treatment;
how many are treated; how eective treatments
are; and what the mental and physical outcomes
of treatment, undertreatment, or no treatment are.
This last point is no exaggeration: the world mental
health surveys showed that in some countries the
prevalence of treatment for severe mental illness is
as low as 2%.
12
We continue to disregard the physi-
cal health needs of people with mental illness and
act as if they are of less worth than others.
13
We now
know that these forms of discrimination can have
lethal consequences.
Competing interests: None declared.
Provenance and peer review: Commissioned; not externally
peer reviewed.
1 Lawrence D, Hancock KJ, Kiseley S. The gap in life expectancy
from preventable physical illness in psychiatric patients in
Western Australia: retrospective analysis of population based
registers. BMJ lu1!;!/6:fl'!9.
l Odegard O. Mortality in Norwegian mental hospitals 19l6-
19/1. Acta Genet Stat Med 19'1;l:1/1-7!.
! Newman SC, Bland RC. Canadian trends in mortality from
mental disorders, 196'-193!. Acta Psychiatr Scand
1937;76:1-7.
/ Lim LC, Sim LP, Chiam PC. Mortality of public mental health
patients: a Singapore experience. Aust N Z J Psychiatry
199!;l7:!6-/1.
' Abiodun OA. Mortality in psychiatric populations: a review.
Cent Afr J Med 1933;!/:lll-'.
6 Ribeiro RB, Melzer-Ribeiro DL, Cordeiro Q. Morbidity and
mortality due to mental disorders in Brazil. Rev Bras Psiquiatr
lu1l;!/:l17-3.
7 Wahlbeck K, Westman J, Nordentoft M, Gissler M, Laursen
TM. Outcomes of Nordic mental health systems: life
expectancy of patients with mental disorders. Br J Psychiatry
lu11;199:/'!-3.
3 Druss BG, Zhao L, Von ES, Morrato EH, Marcus SC.
Understanding excess mortality in persons with mental
illness: 17-year follow up of a nationally representative US
survey. Med Care lu11;/9:'99-6u/.
9 Howard LM, Barley EA, Davies E, Rigg A, Lempp H, Rose D,
et al. Cancer diagnosis in people with severe mental illness:
practical and ethical issues. Lancet Oncol lu1u;3:797-3u/.
1u WHO. Global mental health action plan lu1!-lulu. lu1!.
www.who.int/mental_health/mhgap/consultation_global_
mh_action_plan_lu1!_lulu/en/.
11 United Nations. Convention on the rights of persons
with disabilities. www.un.org/disabilities/convention/
conventionfull.shtml.
1l Wang PS, Aguilar-Gaxiola S, Alonso J, Angermeyer MC, Borges
G, Bromet EJ, et al. Use of mental health services for anxiety,
mood, and substance disorders in 17 countries in the WHO
world mental health surveys. Lancet luu7;!7u:3/1-'u.
1! Thornicroft G. Shunned: discrimination against people with
mental illness. Oxford University Press, luu6.
RESEARCH, p
bmj.com Editorial: Physical health of people with severe mental illness (BMJ luu1;!ll://!)
Despite the right to health having
become law in 126 countries worldwide
(through their ratification of the United
Nations Convention on the Rights of
Persons with Disabilities) it has not
been implemented into effective action
G
O
R
D
O
N

D
I
X
O
N
8 BMJ | 25 MAY 2013 | VOLUME 346
EDITORIALS
bmj.com
OEditorial: Let the patient revolution begin (BMJ lu1!;!/6:fl61/)
OObservations: The future of medicine lies in truly shared decision making
(BMJ lu1!;!/6:fl739)
A call to challenge the Selling of Sickness
A partnership model for a new social health movement
Leonore Tiefer convenor, New View Campaign, Department of
Psychiatry, NYU School of Medicine, New York, NY 1uuu9, USA
ltiefer@mindspring.com
Kim Witczak co-founder, WoodyMatters, Minneapolis, MN
''/u1, USA
Iona Heath immediate past president, Royal College of
General Practitioners, London, UK
The sense that medicine is out of control is gen-
erating a rising tide of concern that includes the
BMJs Too Much Medicine campaign.
1
Throughout
history, unscrupulous people have preyed on our
universal fears of suering and death and made
money by selling dubious remedies. The hope
that the growing scientihc foundation for medical
practice would consign such activity to historical
oblivion has proved to be a vain one. Indeed, con-
temporary enthusiasm for the commercialization
and marketing of healthcare seems to oer ever
wider opportunities to sell medical treatments. The
results of medical research are ohen distorted or
suppressed for commercial gain, and systems that
attempt to control clinicians behaviour through
payment by results drive overdiagnosis and over-
treatment.
2

3
Patients experience well documented
harms as more and more ohen hnancial impera-
tives are allowed to trump clinical judgment.
4

Harm is also caused by well meaning doctors
trying to save lives by diagnosing serious condi-
tions earlier,
5
which inevitably drives up overdiag-
nosis and overtreatment. Well meaning doctors,
patients, politicians, and journalists consistently
overestimate the benehts and underestimate the
harms of most medical treatments.
These problems have been increasingly aired
and analysed on websites, in editorials, and in
dozens of recent books by investigative journal-
ists, crusading physicians, earnest advocates, and
academic reformers.
6
Proposed solutions include
greater regulation of the medical and scientihc
communities to improve accountability and to
expose conicts of interest, greater transparency
in publishing and education, much greater patient
involvement in decision making, and higher stand-
ards of professional integrity. Yet the fundamental
problems remain and the situation seems to get
worse each year.
A comprehensive social health movement
that combines the perspectives and resources
of healthcare professionals, patient representa-
tives, and consumer advocates in a partnership
that aims to counter overdiagnosis and overtreat-
ment is well overdue. A partnership model that
amrms themes of mutual respect and collabo-
ration can enable questions and concerns to be
raised in dierent voices for dierent audiences;
can foster wider dissemination of key messages to
enlarge collective action; and can generate diverse
strategies to promote public awareness and policy
change.
A recent international conference (www.
sellingsickness.com) held in Washington, DC,
brought together many dierent stakeholders in
healthcare, including professionals and consumer
advocates, to discuss how to end disease monger-
ing and how to challenge the undue commercial
inuences that distort so many aspects of health-
care research and practice. This self funded grass-
roots conference deliberately sought to model the
possibility of a new social health movement with
equal representation of professional and advocate
perspectives and extended time for discussion.
A Call to Action on Selling Sickness was devel-
oped by a diverse group over a period of months,
uploaded on to the Selling Sickness conference
website for preview, and opened for signatures
from the hnal day of the conference. Among its
list of concerns were the problems of biased sci-
ence, hidden data, inated diagnostic categories,
unnecessary screening and treatment, and the
widespread neglect of social factors when treat-
ing illness. Its recommendations included a call
for a new movement of alliances and actions to
ensure a hrewall between commercial inuences
and medical guidelines. The campaign aims to
bring an end to direct-to-consumer advertising of
diagnostics and drugs, ensure appropriate testing
of new drugs and devices, put forward reform of
the patent system, and promote responsible health
journalism. Although there was no systematic
eort to obtain signatures from the more than 240
conference attendees, 35 people signed the call for
action by the end of the conference. They repre-
sented nonproht organizations, university research
groups, clinical enterprises, the Public Library of
Science journals, and law omces to name a few,
and came from Canada, the United Kingdom,
Germany, Russia, the Netherlands, New Zealand,
and the United States. The call to action on Sell-
ing Sickness is now online (www.sellingsickness.
com/hnal-statement) and open for individual and
organizational endorsement.
Using professional-advocate collaborations to
tackle the problem of medical overtreatment is not
a new idea. In a recent project, Choosing Wisely,
designed by the American Board of Internal Medi-
cine together with Consumer Reports,
7
medical
specialty organizations identify practices prone to
overuse in their specialties and partner with Con-
sumer Reports to disseminate consumer friendly
educational materials. The list of participating
specialty organizations continues to grow (www.
choosingwisely.com). The alltrials.net movement
has used social media to stimulate publicity and
now has thousands of signatures on its petition for
open access to all clinical trial data (www.alltrials.
net/). The Selling Sickness call to action, promoted
by an emerging advocate-professional partnership,
will add strength to the new social movement for
healthcare reform that may prove crucial to global
health in the 21st century.
peer reviewed.
References are in the version on bmj.com.
SeIIing Sickness, 2013: PeopIe Before Prots
February 21 - 22, 2013
bringing together
ACADEMC MEDCAL REFORMERS,
CONSUMER ACTVSTS, AND HEALTH JOURNALSTS

to examine the current scope of disease mongering
and to deveIop strategies and coaIitions
FOR CHANGE
Hyatt Regency Washington on Capitol Hill
Washington D.C.
Well meaning doctors, patients, politicians, and journalists consistently
overestimate the ben efits and underestimate the harms of most medical treatments
BMJ | 25 MAY 2013 | VOLUME 346 9
EDITORIALS
bmj.com
Feature:
After war, what next?
(BMJ 2010;341:c6910)
Suicide among Falkland war veterans
Responses should be based on sound statistics not misleading sound bites
Defence Statistics (Health) has been monitoring,
analysing, and publishing, the numbers of suicides
in this group and in military personnel from the
same period who did not serve in the Gulf for many
years. These data, reported annually, clearly show
no signihcant excess of suicides directly attribut-
able to service in the 1991 Gulf war.
It is important to note that veterans do seem
to dier from the general population regarding
cause of death in the hrst few years aher serving
in a conict. In the hrst hve years aher discharge,
Vietnam veterans had a slightly increased risk of
all cause mortality compared with the general US
population. This higher risk seems to be mainly
due to external causes of death (including road
tramc incidents).
20
A similar, but non-signihcant,
pattern is seen in British veterans of the Gulf war.
This may reect an inability in a subpopulation of
veterans to adjust to the stresses of life back home
and the relative risks of day to day activities. Vet-
erans may have become used to risky situations
and continue to exhibit risky behaviour in civilian
life, which would make them more prone to being
involved in road tramc incidents. This is consist-
ent with a reported increase in risky driving behav-
iours, such as not wearing a seatbelt, seen in UK
forces deployed in Iraq.
22
We may never know how and why the mean-
ingless and misleading sound bite about absolute
numbers of suicides among Falkland veterans
arose. However, it rapidly caught the popular imagi-
nation of press and public alike, encapsulating a
growing sense of disquiet and unease about the
treatment of ex-service personnel, and this even-
tually led to the generation of the Military Covenant
(which outlines the nations responsibilities to our
Armed Forces personnel).
It is right that we should worry about the longer
term social and psychological consequences of
military service.
2324
This concern should not dimin-
ish, but it should be based on sound statistics, not
catchy sound bites.
Competing interests: NTF is funded by the Ministry of Defence;
SW is the civilian adviser in psychiatry to the British Army and he
is a trustee of Combat Stress; KH and JS are both employed by the
Ministry of Defence. The authors have not been directed by the
Ministry of Defence in any way with regard to this editorial.
peer reviewed.
References are in the version on.bmj.com.
J Holmes medical student, Kings School of Medicine, Kings
College London, London, UK
N T Fear reader in epidemiology, Kings Centre for Military Health
Research, Kings College London, London, UK
K Harrison head of Defence Statistics (Health), Ministry of
Defence (MOD) Defence Statistics (Health), MOD Abbey Wood
North, Bristol, UK
J Sharpley defence consultant adviser in psychiatry,Defence
Medical Services, Ministry of Defence, Department of Community
Mental Health, HM Naval Base, Portsmouth PO LT, UK
S Wessely chair of psychological medicine, Kings Centre for
Military Health Research, Kings College London, London, UK
simon.wessely@kcl.ac.uk
The media sound bite, more Falklands veterans
have taken their own life than were killed in the
war has been in common use for about a dec-
ade. A recent report from the Ministry of Defence
Defence Statistics (Health) has shown it to be
incorrect95 suicides have been recorded among
Falklands veterans, whereas 237 deaths among
UK Armed Forces personnel occurred during
the campaign.
1
The report also found no overall
increase in the suicide rate of Falklands veterans
compared with the general population of the
United Kingdom. So where did this false statistic
come from, and why was it so misleading?
The claim that more Falklands war veterans have
committed suicide than were killed in action was
hrst made in 2002, the 20th anniversary of the
war. A report in the Mail on Sunday gave a hgure
of 264 suicides compared with the accepted hgure
of 237 UK military personnel killed
in action (plus 18 deaths among
coalition forces).
2
The source was
reported as the South Atlantic Medal
Association, the principal group for
UK Falkland veterans, although the
association denied this. Only six months earlier
the Guardian had reported a hgure of around 100
suicides among Falkland veterans.
3
However, the sound bite became ubiquitous,
with the hgure of 264 suicides appearing consist-
ently over a decade in a wide range of media out-
lets,
2

4- 9
peer reviewed scientihc journals,
10
online
blogs,
11

12
political party conferences,
13
and even
a popular trivia book.
14
Nobody seems to have
pointed out that more suicides must have occurred
in this group during the most recent decade. And
neither has anybody mentioned that the compari-
son between the absolute number of suicides and
those killed in action is meaningless because it does
not take into account the number of suicides that
would be expected in the population in general
(irrespective of veteran status).
But the Falklands veterans are not alone in their
falsely reported high risk of suicide. During the
Vietnam war, 58 220 US service personnel were
killed.
15
The hrst time a claim was made that the
number of Vietnam veteran suicides exceeded
that number seems to have been in 1980, in a
manual on the treatment of post-traumatic stress
disorder.
16

17
This number was soon picked up by
the media and soon increasedfor example, the
claim that more than 100 000 veterans committed
suicide was made by a university lecturer in a net-
work television news broadcast.
18
Although there
is no complete population sample or appropriate
control group for all Vietnam veterans, several
good studies have calculated the most likely true
number of suicides among large representative
samples of Vietnam veterans.
16

19

20
As expected,
the number of suicides has steadily increased over
time. For example in 1990, the estimate was 8941
suicides,
16
whereas the most recent estimate, which
used rates calculated in 2004, suggests that the hg-
ure is around 20 000.
20
Thus it is incorrect to repeat,
as many do, that more Vietnam veterans have died
from suicide than were killed in the war.
An eect of service in Vietnam on suicide, even
a substantial one, could possibly be concealed
because the correct comparison is between the
true rate of suicide and the rate of
suicide in people who have not
served in Vietnam. What such stud-
ies show is that there was an initial
non-signihcant rise in the rate of sui-
cide among Vietnam veterans in the
hrst hve years aher discharge. However, over the
full course of the follow-up, suicide mortality rates
were lower than expected compared with the gen-
eral male population in the United States, although
not signihcantly so.
19

20
Similar monitoring of suicide rates has occurred
in the UK in those who served in the 1991 Gulf war.
Forty seven soldiers were killed in combat. By 31
December 2012, 197 veterans had taken their own
lives according to the latest published statistics
21
more than four times the number killed in combat.

Yet there do not seem to have been any claims that
more than four times the number of Gulf veterans
killed in the war, have taken their own lives, which
in this case would be correct. Ministry of Defence
Where did this false
statistic come from,
and why was it so
misleading?
10 BMJ | 25 MAY 2013 | VOLUME 346
EDITORIALS
Preventing admission of older people to hospital
No evidence that managing frail older people in the community reduces admissions
Shaun DSouza specialist registrar
shaun.dsouza@pbh-tr.nhs.uk
Sunku Guptha consultant physician, Medicine for older people,
Peterborough and Stamford Hospitals NHS Foundation Trust,
Peterborough PE GZ, UK
The hypothesis that better community facilities for
older people (over 65 years) will reduce demand
for inpatient care is appealing and has gained
wide acceptance. In England, many commission-
ing groups plan to redirect a substantial proportion
of resources into services that will enable medical
care for older people in the community.
In July 2012, the then secretary of state for health
in England, Andrew Lansley, hrst admitted that
the shih of focus towards community healthcare
would lead to a reduction in inpatient beds. Plans
to reduce beds in medical wards, and in units dedi-
cated to older people, rest on the assumption that
longer term conditions can be safely treated in the
community. Reducing emergency hospital admis-
sions remains a key target, not only because of the
high costs of emergency admission compared with
other forms of care, but also because emergency
admissions disrupt elective healthcare.
1
But will
better community services for older people really
reduce acute admissions to hospital?
In 2010 a paper commissioned by the think tank
the Kings Fund, on avoiding hospital admissions
in all age groups, found no evidence that case man-
agement, in the community or in hospital, reduces
generic admissions.
2
Community matron or case
worker models using experienced nurses targeting
naive and known users of primary or secondary
care in those over 65 years has had no impact on
rates of admission.
3

4
It is hoped that advances in technology, and
the removal of artihcial barriers between hospital
and community care, will oset any reduction in
inpatient capacity. The formation of integrated care
organisations is intended to help break down the
barriers. These organisations emphasise that most
people would prefer to have medical care at home,
and that frail older people can be safely treated
without the need for hospital admission. How-
ever, trials using integrated teams in various forms
have so far failed to reduce numbers of admissions
compared with usual care. One trial targeted low
income older patients in general, but specihcally
those at risk of using secondary care resources. It
used a community multidisciplinary team model
comprising primary care doctors, geriatricians,
community nurses, pharmacists, physiothera-
pists, and social workers. It also involved extensive
assessments and detailed comprehensive manage-
ment plans tailored to individual patients, weekly
meetings to ensure implementation, at least once
monthly patient contact by a team member, and
annual review of protocols by the whole team.
Despite being so thorough and resource intensive,
the intervention had no impact on admissions to
secondary care compared with usual care.
5

Another randomised trial actively screened
community dwelling older people (irrespective of
their contact with primary or secondary care) for
conditions such as depression, falls, urinary incon-
tinence, and cognitive and functional impairment
(the so called geriatric giants). The researchers then
intervened intensively using specialist services that
included geriatric medicine and psychiatry, urol-
ogy, audiology, rehabilitation, psychology, and
social services. However, they found no reduction
in admissions compared with the usual care group
over a three year follow-up.
6
Models of extending secondary care into the
community have failed to reduce admissions
compared with usual care. One trial investigated
a hospital based team consisting of a geriatrician,
trained nurses, and social workers that oered
outreach in the community. Despite active inter-
vention, extensive assessments, and round the
clock support during follow-up, admissions were
not reduced compared with usual primary care.
7

In another trial of older people admitted to hospital
by a geriatric team, intervention during inpatient
stay and extension of follow-up aher discharge did
not reduce admissions compared with usual care.
8
Furthermore, no convincing evidence exists
that increases in the provision of community serv-
ices reduces length of stay for frail older people. A
Cochrane review of hospital at home services
was unable to pool data on length of stay for older
patients admitted with a mix of medical conditions
because of substantial heterogeneity.
9
There is also
no clear association between investment in social
care and hospital bed use among older people.
10
All trials examining community care used the
comprehensive geriatric assessment tool for screen-
ing and intervention, either in its complete or modi-
hed form. This tool has been shown to improve
quality and reduce mortality in older people admit-
ted to secondary care, but it has not been shown to
reduce readmissions, including when instigated at
an early stage in acute medical units or emergency
departments.
11

12
Currently, no validated tool has
been proved to be useful in reducing admissions or
readmissions among frail older people.
Given this lack of evidence, it is difficult to
understand the widespread belief that admissions
to hospital can be reduced by managing frail older
people more eectively in the community. Frailty is
a dynamic condition that is characteristically asso-
ciated with sudden profound decompensation sec-
ondary to a stressor.
13
Aher such events, patients
require dedicated nursing because haemodynamic
instability, hyperactive or hypoactive delirium, and
falls are common. There is no evidence yet that
proactive management in the community can
reduce the occurrence of such episodes, most of
which result in hospital admissions that would be
hard to categorise as avoidable.
We urgently need aids that can more emciently
detect and measure the severity of frailty in com-
munity dwelling older people. These would allow
research into interventions that could reduce epi-
sodic decompensation in older patients, and the
development of integrated care models that are
appropriately directed.
9
There are some known benehts of good commu-
nity care services that speak for additional invest-
ment in these services. Systematic reviews of home
based interventions, despite being complicated by
methodological variations and a lack of stand-
ardisation of interventions, show that community
based patient centred care delivered comprehen-
sively in a sustained fashion with multiple visits
reduces long term institutional care.
14

15
High levels
of patient satisfaction
3-8
and improvements in self
reported health and mental health were found.
16
However, there is no evidence that enhancing
community care for frail older people will reduce
hospital admissions, and demands on second-
ary care will probably continue to rise. There has
been a sustained reduction in the number of acute
beds over the past few decades, and most hospi-
tals now average around 90% bed occupancy.
17
A
further reduction in beds based on the vain hope
that enhancing community services will reduce
admissions could be potentially dangerous to
patient care. It would be more sensible to evaluate
the eects of enhancing community services before
making decisions to cut more acute care beds.
Provenance and peer review: Not commissioned; externally
peer reviewed.
The many benefits of good community care dont
include fewer admissions
BMJ | 25 MAY 2013 | VOLUME 346 11
The BMJ is an Open Access journal. We set no word limits on BMJ research articles, but they
are abridged for print. The full text of each BMJ research article is freely available on bmj.com
Scan this image with your
smartphone to read our
instructions for authors
RESEARCH
RESEARCH NEWS
RESEARCH NEWS All you need to read in the other general medical journals Alison Tonks, associate editor, BMJ atonks@bmj.com
1 year old (more than 80% concordance). Post-
mortem imaging tended to miss infectious and
inammatory causes of death such as sepsis and
myocarditis in older children (concordance 15/28;
53.6%, 35.8 to 70.5).
Autopsy rates are at an all time low in many
countries. We will always need them in some
cases, says a linked comment, but there are clearly
alternatives that mimic more closely the kind
of diagnostic process that goes on before death
(doi:10.1016/S0140-6736(13)60890-9). In this
study, state of the art scans were supplemented by
external examination, premortem and postmortem
blood tests, other non-invasive laboratory tests and
imaging, and an examination of the placenta, but
not by histology. The minimally invasive proto-
col represents an important new collaboration
between pathologists and radiologists that could
have replaced conventional autopsy in an esti-
mated 165 of these 400 deaths.
Lancet lu1!; doi:1u.1u16/Su1/u-67!6(1!)6u1!/-S
Oxandrolone no better than placebo
for severe pressure ulcers
Anabolic steroids are thought to promote wound
healing and can be used to treat some severe
burns. Oxandrolone, however, did nothing for
chronic pressure ulcers in a trial of US veterans
with longstanding spinal injuries. Participants
given the steroid were no more likely to heal
over 24 weeks than controls given a matching
placebo. They all had standard care at one of 16
spinal injury centres run by the US Department of
V eterans Aairs.
Participants, who were almost all men, had seri-
ous and hard to treat pressure ulcers in the sacral,
ischial, or trochanteric regions. Nearly two thirds
had full thickness ulcers. Oxandrolone didnt seem
to help these ulcers heal or prevent recurrence in
the minority that did heal. Eight weeks aher the
end of treatment, 16.7% (18/108) of the oxan-
drolone group and 15.4% (16/104) of controls
had a fully healed ulcer.
The results were a disappointment, and a data
monitoring committee stopped the trial well before
target recruitment. Oxandrolone caused signih-
cantly more abnormal liver function tests than
placebo; it also caused signihcantly more anaemia,
which the authors found harder to explain.
Ann Intern Med lu1!;1'S:71S-l6]
Mycoplasma pneumoniae infection
or harmless colonisation?
Mycoplasma pneumoniae is thought to be an
important cause of respiratory tract infection
in children. Asymptomatic carriage is also com-
mon, according to an epidemiological study from
the Netherlands. A hhh of children admitted to
one hospital for routine surgery had M pneumo-
niae DNA in pharyngeal or nasopharyngeal sam-
ples. So did 16% of 321 children admitted with
symptomatic respiratory tract infections. Stand-
ard diagnostic tests were unable to distinguish
reliably between the symptomatic and asymp-
tomatic children. More than half of both groups
had at least two viral or bacterial pathogens in
their upper respiratory tract.
So the presence of M pneumoniae does not nec-
essarily indicate active infection, even in children
with symptoms, say the authors. We will have to
hnd more eective diagnostic strategies. Children
thought to have clinical infections are ohen given
macrolide antibiotics, and resistance is becoming a
serious problem in some countries. Accurate identi-
hcation of the causative organism, as distinct from
harmless carriage, may help control use of mac-
rolides for this indication, they write.
Bigger and more diverse studies should now be
done to help characterise the pattern and deter-
minants of asymptomatic carriage of M pneumo-
niae, say the researchers. In this study, prevalence
ranged from 3% in the spring of 2009 to 58% in the
summer of 2010.
PLoS Med lu1!;1u:e1uu1///
Preliminary success with combination
treatment for osteoporosis
Treatments for osteoporosis either curb resorption
or encourage formation of bone. Combining the two
approaches doesnt always work, but researchers
from the US report preliminary success with the
antiresorptive antibody denosumab combined with
teriparatide, a derivative of parathyroid hormone
that promotes bone formation.
In an open label trial, postmenopausal women
given both for 12 months had greater improve-
ments in bone mineral density (BMD) at three sites
than controls given either agent alone. The combi-
nation made the biggest dierence at the lumbar
spine (9.1% improvement v 6.2% in women treated
with teriparatide and 5.5% in those treated with
denosumab; P<0.05 for both comparisons). Amgen
and Eli Lilly contributed funding.
The 94 women analysed had a mean age of 66
years and a high risk of fractures. They were not
dehcient in vitamin D. They had daily subcutane-
ous injections of teriparatide, twice yearly subcuta-
neous injections of denosumab, or both.
This is encouraging news for women with osteo-
porosis, says a linked comment (doi:10.1016/
S0140-6736(13)60984-8), although its early days.
The trial wasnt designed to look for an impact on
osteoporotic fractures and doesnt tell us very much
about long term safety and eectiveness of combi-
nation treatment. Teriparatide can be given only
for two years under its current licence. We need to
hnd out what happens when treatment has to stop.
Lancet lu1!; doi:1u.1u16/Su1/u-67!6(1!)6uS'6-9
MRI autopsies look accurate for
infant and fetal deaths
Researchers from two large hospitals in the UK
have developed and tested a minimally invasive
autopsy for childhood and fetal deaths based on
postmortem magnetic resonance imaging (MRI). In
a consecutive series, the minimally invasive option
accurately identihed the same cause of death as
conventional autopsy in 357/400 cases (89.3%,
95% CI 85.8 to 91.9). It worked best for fetal deaths
(more than 90% concordance with conventional
autopsy) and deaths in babies and infants under
Adapted from Lancet 2u13; doi:1u.1u16/ Su1/u-6736(13)6u856-9
Time (months)
C
h
a
n
g
e

i
n

B
M
D

(
%
)
u 3 6 9 12
u
2
/
6
8
1u
C
h
a
n
g
e

i
n

B
M
D

(
%
)
u
1
2
3
/
5
Percentage changes in bone mineral density
Posterior-anterior lumbar spine
Femoral neck
Teriparatide
Denosumab
Both
12 BMJ | 25 MAY 2013 | VOLUME 346
RESEARCH
1
Centre de recherche du CHU de
Qubec, Facult de mdecine,
Universit Laval, Quebec, QC,
Canada
l
Public Health Agency of Canada,
Ottawa, ON, Canada
!
Dalla Lana School of Public Health,
University of Toronto, Toronto, ON,
Canada
/
Population Studies Division, Health
Canada, Ottawa, ON, Canada
'
Department of Epidemiology and
Community Medicine, University of
Ottawa, Ottawa, ON, Canada
6
Clinical Epidemiology Program,
Ottawa Hospital Research Institute,
Ottawa, ON, Canada
Correspondence to: E Lavigne,
Unit de recherche en sant des
populations (URESP), Centre de
recherche du CHU de Qubec,
Hpital du Saint-Sacrement,
1u'u Chemin Sainte-Foy, Quebec,
QC, Canada G1S /LS
elavigne@uresp.ulaval.ca
doi: 1u.11!6/bmj.fl!99
This is a summary of a paper that
was published on bmj.com as BMJ
l1u!;!/6:fl!99
STUDY QUESTION
Does a difference exist in the stage distribution and
post-diagnosis survival among women diagnosed as
having breast cancer between those who have previously
received breast implants for cosmetic purposes and those
with no implants?
SUMMARY ANSWER
The accumulating evidence suggests that cosmetic
breast implants adversely affect breast cancer specific
survival following the diagnosis of such disease.
WHAT IS KNOWN AND WHAT THIS PAPER ADDS
Breast implants are radio-opaque at mammography,
impairing the visualisation of breast tissue and raising
the concern that they may impair the ability to
identify breast cancer at an early stage when survival
is generally more favourable. On the basis of studies
published to date, cosmetic breast augmentation seems
to adversely affect the survival of women
who are subsequently diagnosed as having
breast cancer.
Selection criteria for studies
We did a systematic search of the literature published
before September 2012 in Medline, Embase, Global
health, CINAHL, IPAB, and PsycINFO. Eligible publica-
tions were those that included women diagnosed as
having breast cancer who had had augmentation mam-
maplasty for cosmetic purposes. Two meta-analyses
evaluated whether the stage distribution among women
diagnosed as having breast cancer diered between those
who had received breast implants for cosmetic purposes
and those with no implants and whether cosmetic breast
augmentation before the detection of breast cancer was a
predictor of post-diagnosis survival.
Primary outcome(s)
The primary outcomes were stage distribution of breast cancer
at diagnosis as dehned by non-localised (regional and distant)
versus localised stage and breast cancer specihc mortality.
Main results and role of chance
The overall odds ratio of our hrst meta-analysis based on
12 studies was 1.26 (95% conhdence interval 0.99 to 1.60;
P=0.058; I
2
=35.6%) for a non-localised stage of breast cancer
at diagnosis comparing women with breast cancer who had
implants and women with breast cancer who did not have
implants. The second meta-analysis, based on hve studies,
evaluated the relation between cosmetic breast implantation
and survival. This meta-analysis showed reduced survival
aher breast cancer among women who had received implants
compared with those who had not (overall hazard ratio for
breast cancer specihc mortality 1.38, 1.08 to 1.75).
Bias, confounding, and other reasons for caution
These hndings should be interpreted with caution, as some
studies included in both meta-analyses did not adjust for
potential confounders such as age and period of diagno-
sis. In addition, the meta-analysis on survival included a
relatively small number of studies. Misclassihcation biases
within each study could also be a factor aecting study spe-
cihc measures of association and consequently our pooled
eect. Although we have evaluated the quality of the stud-
ies by using an assessment scale, no threshold scores were
available to distinguish between good and poor qual-
ity studies, which could limit our results as we may have
included studies of poorer quality in our analyses.
Study funding/potential competing interests
This work was supported through scholarship grants by the
Unit de Recherche en Sant des Populations, Cancer Care
Ontario, and the Public Health Agency of Canada.
Breast cancer detection and survival among women with
cosmetic breast implants: systematic review and meta-analysis
of observational studies
Eric Lavigne,
1 l
Eric J Holowaty,
!
Sai Yi Pan,
l
Paul J Villeneuve,
! /
Kenneth C Johnson,
'

Dean A Fergusson,
6
Howard Morrison,
l
Jacques Brisson
1
Association between cosmetic breast implants and breast cancer specic survival
Birdsell et al 1993
/u
Deapen et al 2uuu
/1

Hlmich et al 2uu3
39
Handel et al 2uu7
8
Lavigne et al 2u12
18
Total: P=u.5u1, I
2
=u.u%
u.5 1 2.5
Study
u.9u (u./8 to 1.68)
2.u5 (u.39 to 1u.8u)
1.5/ (u.55 to /.33)
1.81 (1.12 to 2.92)
1.32 (u.9/ to 1.83)
1.38 (1.u8 to 1.75)
Hazard ratio
(% CI)
Hazard ratio
(% CI)
1/.8u
2.11
5./7
25.29
52.33
1uu.uu
Weight (%)
(random eects
analysis)
bmj.com Oncology updates from BMJ Group are at bmj.com/specialties/oncology
BMJ | 25 MAY 2013 | VOLUME 346 13
RESEARCH
STUDY QUESTION What is the life expectancy of people
with mental illness in Western Australia compared with the
general population, and how has this changed over time?
SUMMARY ANSWER The life expectancy gap between
people with mental illness and the general population in
Western Australia, -, increased for males from
. to . years and for females from . to . years.
WHAT IS KNOWN AND WHAT THIS PAPER ADDS People
with mental illness have a shorter life expectancy than the
general population. In Western Australia the gap increased
between and , and the majority of excess
mortality in people with mental illness was attributed to
common physical health conditions such as heart disease,
respiratory disease, and cancer.
Participants and setting
Our study was based on administrative registers describing
psychiatric patients and the general population of Western
Australia, 1985-2005.
Design
We used a population based register of contacts with
mental health services, including inpatient, outpatient,
and community mental health clinics contacts. Using
record linkage we calculated mortality rates of psychiatric
patients and from these we calculated life expectancies.
Primary outcomes
The study outcomes were life expectancy of people with
mental illness, by diagnosis, from 1985 to 2005 compared
with life expectancy of the general population, and the pro-
portion of excess deaths attributed to each major cause
of death.
Main results and the role of chance
In the general population, life expectancy in males
increased from 73.1 years in 1985 to 79.1 years in 2005,
and in females from 79.3 years to 83.8 years. In psychiat-
ric patients, life expectancy in males increased from 59.6
years (95% conhdence interval 58.8 to 60.3) to 63.2 years
(62.6 to 63.7) and in females from 68.9 years (68.1 to
69.6) to 71.8 years (71.2 to 72.4). The life expectancy gap
between the general population and psychiatric patients
widened from 13.5 years (12.7 to 14.3) to 15.9 years (15.3
to 16.5) for males and from 10.4 years (9.6 to 11.2) to 12.0
years (11.3 to 12.6) for females between 1985 and 2005.
Additionally, 77.7% of excess deaths were attributed to
physical health conditions, including cardiovascular dis-
ease (29.9%) and cancer (13.5%). Suicide was the cause
of only 13.9% of excess deaths.
The study was based on administrative data relating to
people in contact with mental health services. People
with undiagnosed or untreated mental health problems
or people only treated by general practitioners were not
covered. People with mental disorders who were not in
contact with services may have had dierent, and possibly
worse, mortality outcomes. Changes in life expectancy over
time could be inuenced by changes in service delivery and
diagnostic practices. However, as the prevalence of contact
with mental health services has increased over time this
would be expected to reduce, not increase, the observed
gap in life expectancy.
Generalisability to other populations
Western Australia has a similar mental healthcare system
to other parts of Australia, many European countries, New
Zealand, and Canada. Although the generalisability of the
increasing gap in life expectancy is unknown, the size of
the gap and main contributing causes are likely to be simi-
lar in other locations.
This study was supported by a grant from the Grimth Insti-
tute for Health and Medical Research. The funding source
had no role in the conduct of the study or its publication.
The gap in life expectancy from preventable physical illness
in psychiatric patients in Western Australia:
retrospective analysis of population based registers
David Lawrence,
1
Kirsten J Hancock,
1
Stephen Kisely
l !
1
Telethon Institute for Child Health
Research, Centre for Child Health
Research, The University of Western
Australia, PO Box S'' West Perth
WA 6S7l Australia
l
School of Population Health, The
University of Queensland, Brisbane,
Australia
!
Griffith Institute for Health
and Medical Research, Griffith
University, Brisbane, Australia
Correspondence to: D Lawrence
dlawrence@ichr.uwa.edu.au
doi: 1u.11!6/bmj.fl'!9
lu1!;!/6:fl'!9
OEDITORIAL by Thornicroh
Year
L
i
f
e

e
x
p
e
c
t
a
n
c
y

Life expectancy of people with mental illness compared with

general population of Western Australia, by year and sex
Men
L
i
f
e

e
x
p
e
c
t
a
n
c
y
Women
% CI People with mental illness
Western Australia population
bmj.com
OResearch: Mental disorders
and vulnerability to homicidal
death
(BMJ lu1!;!/6:f''7)
OResearch: Association
between psychological distress
and mortality
(BMJ lu1l;!/':e/9!!)
OResearch: Mortality aher
hospital discharge for people
with schizophrenia or bipolar
disorder
(BMJ lu11;!/!:d'/ll)
14 BMJ | 25 MAY 2013 | VOLUME 346
RESEARCH
Completeness and diagnostic validity of recording acute
myocardial infarction events in primary care, hospital care, disease
registry, and national mortality records: cohort study
Emily Herrett,
Anoop Dinesh Shah,
Rachael Boggon,

Spiros Denaxas,
Liam Smeeth,

Tjeerd van Staa,

Adam Timmis,
Harry Hemingway
STUDY QUESTION How do multiple linked data sources

from primary care, hospital care, disease registry, and death
records compare for recording of fatal and non-fatal acute
myocardial infarction?
SUMMARY ANSWER Each data source missed a substantial
proportion of myocardial infarction events (between %
and %). This incomplete ascertainment means that
incidence based on a single source is lower than using any
combination of sources.
WHAT IS KNOWN AND WHAT THIS PAPER ADDS Electronic
health records are increasingly used in research for
measuring outcomes of healthcare and in health policy, but
no studies have addressed the completeness and validity
of recording of myocardial infarction across four national
health record sources. Using linked multiple sources can
help to overcome the incomplete ascertainment that occurs
when relying on a single data source.
Our study was based on a sample of patients in England
within the Clinical Practice Research Datalink who had an
acute myocardial infarction recorded between 2003 and 2009
in one of four linked data sources: Clinical Practice Research
Datalink (primary care data), Hospital Episode Statistics (hos-
pital admissions), the disease registry MINAP (Myocardial
Ischaemia National Audit Project), and the Omce for National
Statistics mortality register (cause specihc mortality data).
Design, size, and duration
We identihed 21 482 patients with a record of fatal or non-
fatal acute myocardial infarction. Once we identihed an
acute myocardial infarction in any one source, we exam-
ined the remaining sources for their agreement in the diag-
nosis and timing of records.
The patients identihed in each of the sources were compa-
rable for age and sex distribution and prevalence of cardio-
vascular disease risk factors. 31.0% of patients with non-fatal
acute myocardial infarction were identihed in all of primary
care, hospital admissions, and disease registry sources, and
63.9% in two or more sources. This was reected in incidence
estimates from each source, which were lower when using
one source to ascertain cases than when using all sources
combined. Younger, male patients with a lower rate of pri-
mary care consultations were more likely to be recorded in
multiple sources. Fatal infarcts were likely to be recorded in
primary care and in mortality statistics, but much less likely
to be recorded in hospital admissions or the disease registry.
These data were from a sample of 244 English general
practices that consented to linkage. In terms of age and
social deprivation, however, they were representative of
all general practices in England. Recording standards
in included practices may be higher than those in non-
included practices and therefore agreement across all
English practices may be lower than described.
Generalisability to other populations
These hndings are relevant to all countries that record
acute myocardial infarction in multiple dierent electronic
health record sources.
This work was supported by grants from the UK National
Institute for Health Research (RP-PG-0407-10314), the
Wellcome Trust (086091/Z/08/Z), and UK Biobank.
LS is supported by a senior clinical fellowship from the
Wellcome Trust (098504). EH is supported by a Medical
Research Council studentship. AS is supported by a clini-
cal research training fellowship from the Wellcome Trust
(0938/30/Z/10/Z). Clinical Practice Research Datalink is
owned by the UK Department of Health and operates within
the Medicines and Healthcare products Regulatory Agency.
(For full details see bmj.com.)
London School of Hygiene and

Tropical Medicine, London
WCE HT, UK
Department of Epidemiology and

Public Health, Clinical Epidemiology
Group, University College London,
UK
Clinical Practice Research Datalink

Group, Medicines and Healthcare
products Regulatory Agency,
London, UK
Utrecht Institute for Pharmaceutical

Sciences, Utrecht University,
Utrecht, Netherlands
Barts and the London School of

Medicine and Dentistry, London, UK
Correspondence to: E Herrett
emily.herrett@lshtm.ac.uk
doi: ./bmj.f
;:f
Disease registry
myocardial
infarction
Hospital data
myocardial
infarction
Primary care
myocardial
infarction
1/88
(8%)
18u6
(1u%)
3188
(18%)
3532
(2u%)
5561
(31%)
1u99
(6%)
129u
(7%)
Number and percentage of records recorded in primary
care (Clinical Practice Research Datalink), hospital care
(Hospital Episode Statistics), and disease registry
(Myocardial Ischaemia National Audit Project) for
non-fatal myocardial infarction across the three sources
(n= patients)
BMJ | 25 MAY 2013 | VOLUME 346 15
RESEARCH
STUDY QUESTION Does measurement of C reactive
protein (CRP) and procalcitonin help in the diagnosis of
pneumonia in primary care?
SUMMARY ANSWER CRP concentration at the optimal
threshold of > mg/L increases diagnostic certainty in
the patients in whom diagnostic doubt remains after
history and physical examination, while procalcitonin
adds no clinically relevant information.
WHAT IS KNOWN AND WHAT THIS PAPER ADDS There is
limited evidence on the diagnostic accuracy of signs and
symptoms for pneumonia that is applicable to primary
care, and the usefulness of additional measurement of
CRP and procalcitonin is largely unknown. A clinical rule
based on symptoms and signs to predict pneumonia in
patients presenting to primary care with acute cough
performs best in patients with mild or severe clinical
presentation. Addition of CRP at the optimal cut off of >
mg/L improves diagnostic certainty but measurement of
procalcitonin adds no clinically relevant information.
Adult patients presenting with acute cough in 16 primary
care centres in 12 European countries.
Design, size, and duration
Between October 2007 and April 2010, 2820 patients had
their history taken, underwent physical examination and
measurement of C reactive protein (CRP) and procalci-
tonin in venous blood on the day they hrst consulted,
and underwent chest radiography within seven days.
Pneumonia was dehned as present if the local radiolo-
gist recorded lobar or bronchopneumonia.
Of the 2820 patients (mean age 50, 40% men), 140 (5%)
had pneumonia. The optimal combination of history and
physical examination for diagnosis included absence of
runny nose and presence of breathlessness, crackles and
diminished breath sounds on auscultation, tachycardia,
and fever, with the area under the receiver operating
characteristic curve of 0.70 (95% conhdence interval
0.65 to 0.75). Addition of CRP at the optimal cut o of
>30 mg/L increased the area under the curve to 0.77
(0.73 to 0.81). Signs and symptoms were useful in cor-
rectly identifying patients with a low (<2.5%) or high
(>20%) diagnostic risk in 26% of patients. In the 74%
of patients in whom diagnostic doubt remained (esti-
mated risk 2.5%-20%), CRP helped to correctly exclude
pneumonia (net reclassihcation improvement 28%). A
simplihed diagnostic score based on symptoms, signs,
and CRP resulted in pneumonia proportions of 0.7%, 4%,
and 18% in the low, intermediate, and high risk group,
respectively. Procalcitonin had no clinically relevant
added value in this setting.
Many more eligible patients presented during the recruit-
ment period than were approached about participation
in this study and therefore we probably did not recruit all
consecutive eligible patients. Nevertheless, clinical selec-
tion bias is unlikely because feedback from recruiting
clinicians during and aher the study showed that recruit-
ment of every eligible patient was impossible because
of the time required to recruit and assess each patient.
CRP and procalcitonin concentrations were measured
with conventional venous blood tests in a diagnostic
laboratory and not a point of care test. The added value
of CRP might be dierent if measured with a point of care
test in general practice, but other studies have shown
good agreement between such test results and a conven-
tional laboratory test.
Chest radiographs were examined by local radiolo-
gists. We attempted to increase uniformity in assessment
by implementing a protocol for reporting. While interob-
server variability was present, the of 0.45 (moderate
agreement) was comparable with previous studies.
Use of serum C reactive protein and procalcitonin concentrations
in addition to symptoms and signs to predict pneumonia in patients
presenting to primary care with acute cough: diagnostic study
Saskia F van Vugt,
1
Berna D L Broekhuizen,
1
Christine Lammens,
2
Nicolaas P A Zuithoff,
1

Pim A de Jong,
3
Samuel Coenen,
2
Margareta Ieven,
2
Chris C Butler,
4
Herman Goossens,
2

Paul Little,
5
Theo J M Verheij,
1
on behalf of the GRACE consortium
1
University Medical Center Utrecht,
Julius Center for Health Sciences
and Primary Care, PO Box S''uu,
!'uS GA Utrecht, Netherlands
l
University of Antwerp, Laboratory
of Medical Microbiology, Vaccine
and Infectious Diseases Institute
(VAXINFECTIO), Antwerp, Belgium
!
University Medical Center Utrecht,
Department of Radiology, Utrecht,
Netherlands
/
Institute of Primary Care and Public
Health, School of Medicine, Cardiff
University, Cardiff, Wales
'
Primary Care Medical Group,
University of Southampton Medical
School, Southampton, UK
Correspondence to:
B D L Broekhuizen
b.d.l.broekhuizen@umcutrecht.nl
doi: 1u.11!6/bmj.fl/'u
lu1!;!/6:fl/'u
Comparison of diagnostic risk for pneumonia by diagnostic model with and without addition of CRP testing
Risk according to symptoms and
signs model(without CRP)
Risk according to symptoms and signs model plus CRP > mg/L
Patients with pneumonia Patients without pneumonia
<.% .-% >% Total <.% .-% >% Total
<l.'% / (!6)* 7 (6/) u (u) 11 '6S (S7) S6 (1!) u (u) 6'/
l.'-lu% l7 (l6) '6 ('!)* ll (l1) 1u' 9'7 (/S) 966 (/9) 6/ (!) 19S7
>lu% u (u) ' (l1) 19 (79)* l/ u (u) 1l (!1) l7 (69) !9
Total !1 6S /1 1/u 1'l' 1u6/ 91 l6Su
*Patients classified in agreement according to model with and without CRP >!u mg/L. Of all patients with pneumonia, l9 (ll+7+u) are reclassified to higher risk groups and !l (l7+')
to lower risk groups. For patients without pneumonia this is 1'u (S6+6/) and 969 (9'7+1l), respectively. Reclassification improvement is l% among patients with pneumonia (l9-
!l of 1/u) and !u% among patients without pneumonia (9'7-1'u of l6Su), resulting in net reclassification improvement of l+!u=lS% (9'% CI u.17 to u./u).
bmj.com
OResearch: Serum glucose
levels for predicting death in
patients admitted to hospital for
community acquired pneumonia
(BMJ lu1l;!//:e!!97)
16 BMJ | 25 MAY 2013 | VOLUME 346
HEAD TO HEAD
Frances Flinter professor, Department of Clinical
Genetics, Guys and St Thomas NHS Foundation Trust,
London frances.flinter@gstt.nhs.uk
This is an exciting time to work
in medicine. The enormous
advances in our understanding
of inherited diseases that have owed from
sequencing the human genome mean that
many patients beneht from faster, more accurate
diagnoses, targeted screening, cascade testing
for relatives, and the availability of prenatal
diagnosis and pre-implantation genetic
diagnosis. Clinicians see patients, take a family
history, examine them, and then decide which
tests will conhrm or exclude specihc diagnoses.
Sequencing the hrst human genome took
13 years and cost $3bn (E2bn); today we can
sequence a complete human genome in a few
days for a few thousand pounds. The price will
fall further, though the cost and challenges of
analysing, interpreting, and storing the data are
substantial.
1

2
Mutation screening within a single gene is
rapidly being replaced by targeted screening
of a relevant panel of genes known to be
implicated in particular phenotypes, and
whole exome or whole genome sequencing
is likely to become the preferred approach
in these situations over the next few years.
3

Sequencing data can be hltered so that only
variants in relevant genes are analysed; variants
that might cause disease can then be validated by
functional studies before their pathogenicity is
conhrmed. In some patients with cancer, detailed
genetic studies to identify known variants in DNA
extracted from tumours can help to determine
the most appropriate chemotherapyso called
personalised medicine.
Ethical questions
The UK Genetic Testing Network has led the
world in developing a governance framework that
ensures proper evaluation of the clinical validity
and utility of tests before they are used in clinical
practice. The governments announcement in
December that it plans to sequence the genomes
of 100 000 NHS patients in the next three to hve
years is recognition of the wealth of information
that will be revealed. Plans to do this on a research
basis, concentrating on the genomes of people
with cancer and rare diseases, are sensible, as
detailed knowledge of individual phenotypes will
support data interpretation, increasing the library
of reference information available.
4

John Burn professor, Institute of Genetic Medicine,
Newcastle University, UK john.burn@newcastle.ac.uk
In 1986 our 5 year old son
planted a conker in our back
yard. He explained that he
wanted a tree house so needed a tree. The tree is
now ready to receive boarders, though they will
be from our familys next generation. The Human
Genome Project began at about the same time
with similar high aspirations: to deliver personal-
ised medicine to generations to come.
That time is now upon us. The cost of gene
sequencing has fallen 10 000-fold in a decade,
with another drop by an order of magnitude
expected soon. Setting aside the considerable
but surmountable challenges associated with
large segments being duplicated or deleted and
stretches of hard to sequence repeats, we can have
a whole genome for the price of a family package
holiday. Even now, bulk sequencing all 20 000
genes, the exome, costs less than E500.
The Human Genome Project depended on Brit-
ish discovery, particularly the work of the Well-
come Trusts Sanger Institute. No one can match
our capacity to do genetic medicine in a coordi-
nated healthcare system at scale. This is why the
UK government has committed E100m to pump-
prime sequencing of 100 000 whole genomes in
the NHS. Only by analysing sequence data and
phenotypes across large patient populations will
we understand which bits of genetic information
are clinically relevant. We can and should lead the
world in showing how to put genetic knowledge
into patient care safely and eectively.
Responsible use
Genetic predisposition plays a core role in most
common diseases. It is the primary cause of most
of the rare diseases that collectively amict 1 in 17
people.
1
Clinically relevant discoveries are enter-
ing practice at a rate of more than 30 a month.
And the provisos? First do no harm. Our capac-
ity for interpretation is still rudimentary so we
must have explicit consent to retain sequence
data linked to patients records. Our population
trusts the health service to deliver healthcare to
all in need, regardless of genetic predispositions.
We must keep that principle safe along with the
sequence data. That doesnt preclude partner-
ship with the drug and biotechnology industries:
we need them to expand the exciting list of drugs
that can target genetic subgroups and give us the
gadgets and biomarker algorithms to hnd them.
As we leave the high ground of single gene
disease, such as hereditary cancers and cardio-
myopathies, we risk drowning in data and doing
harm. Oering volunteers at risk of monogenic
disorders an eective service in return for them
allowing their genomes to be pooled and data
mined is straightforward. Systematic gathering
Should we
sequence
everyones
genome?
As technological prowess soars
and costs plummet, is the era
of personalised medicine now
in sight? John Burn says that
sequencing everyones genome
would give us unparalleled
knowledge to prevent, diagnose,
and treat disease, but there are
serious ethical implications thinks
Frances Flinter
Possessing the technical ability to do
something new is not an immediate
justification for going ahead with it
bmj.com
Poll: Should we sequence
everyones genome?
Vote at bmj.com
News: Whole genome sequencing fails
to predict risk of most common diseases
(BMJ ;:e)
Editorial: Putting genomics
into practice
(BMJ ;:d)
Sequencing is now cheap, but what are the costs
in analysing, interpreting, and storing the data?
BMJ | 25 MAY 2013 | VOLUME 346 17
HEAD TO HEAD
The PHG Foundation, an independent
think tank on biomedical science, welcomed
this initiative but identihed areas that require
clarihcation before work commences. What will be
the nature of the full and explicit consent given
by patients before their genome is sequenced?
What happens if they change their mind? What
information will be provided about the results and
by whom? How will implications for relatives be
handled? What secondary uses of the data will be
allowed? How will public trust in health services
be maintained in the light of possible commercial
gain by the private sector from using the data
generated?
5
Sequencing every persons genome is a very
dierent proposition. Our extremely limited
knowledge makes meaningful interpretation of
most of the information that would be generated
impossible. On average, a normal, healthy person
carries about 400 potentially damaging DNA
variants plus two variants known to be associated
directly with disease traits.
6
The hgure of 400
is likely to increase as more powerful genetic
studies discover rare variants more emciently. It
is dimcult to predict the clinical consequences of
most of these variants and to prepare individuals
to receive this sort of information.
7
The costs
attached to providing additional infrastructure for
counselling, imaging, and bioinformatics would
be considerable.
8
There are even greater concerns
about how people with mental health disorders or
learning dimculties could be protected from harm.
9
Privacy cannot be guaranteed, even if data
have been entered into supposedly anonymous
genetic databases. A recent study showed that
it was possible to identify some individuals who
had made their genomes available anonymously
to researchers by cross referencing genomic
sequence data, and other data attached to it, with
information available online from genealogy
databases. These exploit the fact that short tandem
repeats on Y chromosomes are highly heritable
from father to son and can be linked to surnames.
10
Knowledge may be unhelpful
People may feel fatalistic if they are told they
have genetic variants that predispose them to
coronary artery disease, such as by making them
less inclined to take steps to protect themselves.
We know that 90% of adults at risk of inheriting
Huntingtons disease choose not to have a
predictive test, preferring not to know their genetic
status. For children there is a strong international
consensus against genetic testing just for
information (as opposed to helping with diagnosis
or management of a condition in childhood). For
most people, simple lifestyle measures such as
a good diet, maintaining a normal weight, not
smoking, not drinking excessively, and exercising
regularly have far more eect on health than any
genetic variants. Yet it seems remarkably dimcult
to communicate these public health messages
eectively.
If we sequence individuals DNA and tell
them that they are genetically predisposed to be
slightly more at risk of common diseases, we may
be doing them a great disservice, demotivating
them from behaving sensibly. And the private
sector will see a marketing opportunity for all
sorts of clinically unnecessary and potentially
damaging screening, with further negative and
unintended consequences.
Possessing the technical ability to do
something new is not an immediate justihcation
for going ahead with it, especially in such an
ethically complex area. Good medical practice
requires tests to answer a specihc question
with a reasonable expectation of results being
interpretable and useful. Currently, whole
genome sequencing in healthy individuals has
nothing to oer clinically because most of the
data generated are meaningless; the maxim hrst
do no harm still holds.
peer reviewed.
of genomic information where there has been no
request, and any sequence variation discovered
lacks clinical use, is more challenging if consent
and follow-up counselling are to be eective.
Each one of us carries perhaps three million
sequence variants, of which about 400 contribute
to disease predisposition,
2
and one or two would
cause a severe recessive disease if both parents
pass them to a child. The bioinformaticians need to
know it all to develop robust diagnostic algorithms.
But patients do not. We must not dump heaps
of molecular uncertainty on patients, families,
and their carers. In Nijmegen, the Netherlands,
teams of geneticists try to interpret exomes of
patients provided by clinicians, passing on (with
explicit consent) only information about variants
of known relevance plus incidental hndings
of obvious importance (H G Brunner, personal
c ommunication).
Personalised treatment
Genomics is not just about rare syndromes and
predisposition to disease. It underpins huge vari-
ation in our capacity to metabolise drugs, ohen
leading to serious adverse events, but this is
ignored. At the moment everybody gets the same
size shoes and they are asked to hobble back next
week if they dont ht. We have known for decades
many of the simple genetic variants responsible
for such adverse events. We write more than half
a million prescriptions for warfarin each year,
knowing that three genetic variants can help
quickly to reach the individual dose, cutting
attendances and adverse events; yet still we just
guess.
3
Widespread sequencing linked to out-
comes will expand such knowledge considerably.
I remain sceptical of an early transition to pro-
viding our genome as part of our electronic medi-
cal records. The sequences we can provide now
are not sumciently complete, and safe storage and
access present challenges.
When it costs only E100 we can run the
sequencers more than once, extracting necessary
information and discarding the rest. In some situ-
ations we will be able to reduce the question to a
genetic test at the point of care.
I am part of a team thats been working for
hve years on using nanowires to analyse nucleic
acids.
4
We are about to test disposable cassettes
that will extract, amplify, and perform specihc
tests such as drug sensitivity in under 15 minutes
for under E20. The next step will be to provide
whole genomes. Others are also innovating in this
held. Whoever wins the race, the strong probabil-
ity is a mixed economy of stored whole genomes,
disposable sequencing in hospitals, and cheap,
fast genotype panels in some frontline care set-
tings. The net result will be accurately targeted
diagnosis, treatment, and prevention.
Genomics extends beyond identifying pre-
disposition to disease. Linking whole genome
sequencing to clinical outcomes will inuence
drug discovery and development. The BRAF
gene inhibitor vemurafenib to treat melanoma
was developed only a decade after the cancer
genome project identihed the target.
5
And the
tools developed from human genomics will
transform another battlefront: routine high speed
sequencing of pathogens will expose their weak-
nesses within hours, revolutionising response to
epidemics worldwide.
Meanwhile point of care technology will
allow drug resistance of infective agents causing
diseases such as malaria to be studied in real time.
Whole genome sequencing is a technical, clinical,
and societal challenge. But as Goethe said, What-
ever you do, or dream you can, begin it. Boldness
has genius and power and magic in it.
Competing interests: I am chair of the British Society for
Genetic Medicine and genetics lead for the National Institute
of Health Research. I am a shareholder and director of
QuantuMDx Group, which is involved in nanowire based
genotyping and sequencing.
peer reviewed.
Genetic predisposition plays a core role in
most common diseases. It is the primary
cause of most of the rare diseases that
collectively afflict 1 in 17 people
18 BMJ | 25 MAY 2013 | VOLUME 346
PSYCHIATRY
Farley, like many critics, is concerned about
DSM-5s contribution to what he calls the relent-
less production of disorders and pathologising of
normal extremes.
6
But it remains true, says Farley, that psychiatry
and psychology in general need to be better con-
nected with the real world. I worry that we have
too much of the monastic science in us; too many
laboratory studies. People dont live in labora-
tories, they dont grieve in laboratories, they dont
feel pain in laboratories.
The American Psychiatric Association, which
has invited criticism and contributions through-
out the long revision process, and since 2010
has received an unprecedented 15 000 com-
ments,
7
has always been vigorous in defending
its position. But its defence of the removal of the
bereavement exclusion, written by a member of
the DSM-5 mood disorder work group, seems to
conhrm wider fears that the associations commit-
tee led system for selecting disorders for inclusion
is a recipe for disorder escalation.
The grief exclusion criteria, wrote Kenneth
Kendler of the Virginia Institute for Psychiatric
and Behavioural Genetics, in a document posted
on the APA website in 2011, had been added to
DSM-III largely on the basis of the work of one
of the DSM-III task force members who was then
studying grief and was carried forward with little
modihcation into DSM-IV.
8
Pet interests
Peter Tyrer, professor of community psychiatry at
Imperial College London and editor of the British
Journal of Psychiatry, is chair of the personality
disorder group for the revision of ICD-11, due
out in 2015, and says his held oers a very good
example of what I think is wrong with DSM. A lot
of clever people sit around a table and say Ive
done work on this and I want to have narcissistic
personality disorder included, I want to have
dissocial personality disorder, I want to have
avoidant personality disorder.
In fact, he says, these are categories of person-
ality disorder that actually have no scientihc basis
behind them and yet here are worthy people sit-
ting in committees all agreeing that this is impor-
tant and weve got to include them. The result,
he says, is that you are in danger of medicalising
people unnecessarily.
I
t isnt every day that a new medical handbook
attracts as much media attention as the latest
blockbuster novel by a world famous author.
But the controversy surrounding the publica-
tion on 22 May of the long awaited hhh edition
of the Diagnostic and Statistical Manual of Mental
Disorders has generated international coverage
on a scale to rival that aorded last weeks release
of the latest sin and symbols thriller from the best-
selling author Dan Brown.
And, though unlikely literary bedfellows,
DSM-5 and Infernoa romp through clues
plucked from Dante Alighieris 14th century poem
Divine Comedyhave something else in common:
the authors of both have been going through hell
at the hands of their critics.
The American Psychiatric Association pro-
duced the first DSM in 1952 and in the inter-
vening 60 years there have been six subsequent
iterations. The last was a text revision of 1994s
DSM-IV, published as DSM-IV-TR in 2000, which
means that DSM-5 (the Roman numerals have
been abandoned for this latest edition in defer-
ence to the demands of the digital age) has been
13 hard fought years in the making, during which
its makers have come under unprecedented scru-
tiny and attack.
For the hrst few decades of its existence, the
DSM had a fairly low public prohle, although one
of the seeds of discontent that has now reached
full bloomthe suggestion that the World Health
Organizations International Classi-
hcation of Diseases (ICD) is the only
manual the world of psychiatry and
psychotherapy now needswas in
fact sown at the birth of DSM in
1952.
In 1948 WHO produced the sixth
edition of the ICD, which for the
hrst time included a category for
mental disorders. Although this
drew heavily on the experience of
the US Veterans Administration
and the classihcations of mental disorders it had
developed while working with second world war
military personnel, the American Psychiatric
Association committee on nomenclature and sta-
tistics decided to produce its own manual.
1

2
DSM hrst came to the attention of the popular
press in 1973, when members of the American
DSM-5: a fatal diagnosis?
Unprecedented levels of criticism have marked the run-up to this
weeks launch of DSM-5. Jonathan Gornall asks whether this
mighty US psychiatric handbook has finally over-reached itself
bmj.com
ONews: More psychiatrists attack
plans for DSM-'
(BMJ lu1l;!//:e!!'7)
OPersonal view: The new somatic
symptom disorder in DSM-' risks
mislabelling many people as mentally ill
(BMJ lu1!;!/6:f1'Su)
Psychiatric Association embarking on the revision
process that would lead to DSM-III began deliber-
ating whether homosexuality should be stricken
from the associations omcial catalog of mental
disorders, in which it was still listed as a sexual
deviation, alongside sadism and masochism.
3
Among those who agreed it should go was
R obert Spitzer, the man who, as chair of the task
force that produced DSM-III in 1980, is widely
credited with having been the architect of the
modern DSM and its classihcation of disorders.
Much later, however, Spitzer seemed to have sec-
ond thoughts about the approach to psychiatry
that DSM-III had crystallisedand the reserva-
tions he expressed six years ago are being echoed
loudly today by critics of DSM-5.
Medicalising ordinary experiences
We made estimates of the prevalence of medi-
cal disorders totally descriptively, he told a BBC
documentary in 2007, without considering that
many of these conditions might be normal reac-
tions which are not really disorders, because we
were not looking at the context in which those
conditions developed.
4
They had, he agreed,
to some extent medicalised much of ordinary
human sadness, fear, ordinary experiences.
This is one of the key criticisms of DSM-5
5
and
few have presented it as eloquently as Frank Far-
ley, the former president of the American Psycho-
logical Association, who takes personal exception
to the widely condemned
decision to remove from
DSM-5 an exclusion clause
that prevents bereavement
being diagnosed as a major
depressive disorder for up to
two months aher the death of
a loved one.
Grieving is normal, he
says, and if you grieve for
more than two months should
you be labelled with some
kind of mental illnessand thats the terminol-
ogy that the public will pick up on? The answer is
clearly no. My hrst wife died three decades ago; I
dont grieve in the usual way that I did for a long
time aher she passed away, but I still think of her
and I still have emotional memories. Nothing
wrong with thatits human nature.
A lot of clever people
sit around a table and
say Ive done work on
this and I want to have
narcissistic personality
disorder included, I
want to have dissocial
personality disorder,
I want to have avoidant
personality disorder
BMJ | 25 MAY 2013 | VOLUME 346 19
PSYCHIATRY
Although he concedes that for a long time the
under-resourced ICD followed in the wake of
DSM, Tyrer is not alone in believing that, while it
has been a stimulus to interest in classihcation,
DSM has over-reached itself and, in the face of a
strengthened, streamlined ICD, is now on the
way out.
I think there are disorders out there that are
pets of particular doctors or groups of d octors,
says Gary Greenberg, a psychotherapist in
Co nnecticut and author of the newly published
The Book of WoeThe DSM and the Unmaking of
P sychiatry.
9
They have an agenda to get [a disorder] into
the DSM and they succeed partly because they are
the people on the committee.
The best example of this in DSM-5, he says,
is disruptive mood disregulation disorder, one
of several new depressive disorders. It has been
introduced, says the APA, to address concerns
about potential overdiagnosis and overtreatment
of bipolar disorder in children.
10
Other newcomers include excoriation disorder,
for which there is strong evidence for its diag-
nostic validity and clinical utility, and hoard-
ing disorder, which reects persistent dimculty
discarding or parting with possessions due to
a perceived need to save the items and distress
associated with discarding them.
10
Curiously, one of the fiercest critics of the
DSM-5 process has been Allen Frances, the chair
of the DSM-IV task force. Since 2010, his less than
friendly hre from the high ground of a column in
Psychology Today has sniped at everything from
the price gouging of DSMs $199 (E130; t155)
cover priceto the suggestion that the manual
will give drug companies running room to con-
tinue their disease mongering of female sexual
disorders.
11

12
He joins a diverse band of opponents, includ-
ing the Society for Humanistic Psychology, which
in January 2012 began a public exchange by
sending an open letter to the DSM-5 task force
asking it to submit its proposals to an independ-
ent group of scientists. The letter was backed by
an online petition that attracted the signatures
of more than 14 000 individuals and 53 organi-
sations, including the British Psychological
S ociety.
13

14
It spoke of a sort of unease, to say the least,
says Farley, and basically we got a thanks, but
no thanks. They said they had the science well
covered in terms of expertise.
Internal dissent
But not everyone on the task force agreed. In April
last year, two members of the DSM-5 personality
orders work group walked o the job, claiming
the members had thrown away an important
opportunity to advance the study of personality
disorder by developing an evidence-based classi-
hcation with greater clinical utility than DSM-IV.
Instead, with a truly stunning disregard for evi-
dence, they had advanced a seriously awed
proposal which was unnecessarily complex,
incoherent, and inconsistent.
Important aspects of the proposal, wrote Roel
Verheul and John Livesley, in an email made pub-
lic by Frances, lacked any reasonable evidential
support of reliability and validity. Not surpris-
ingly, they added, the proposal has received
widespread criticism to which the work group
seems impervious.
Verheul, professor of personality disorders at
the Viersprong Institute in Amsterdam, and Lives-
ley, former head of psychiatry at the University
of British Columbia, were the only international
members of the group, which Frances dismissed
as a small group of cloistered DSM 5 experts
stubbornly ignoring the sharp criticism from
within their own group and the near universal
rejection of their proposals by everyone else in
the held.
15
Competing interests
One of the most serious accusations levelled
against DSM is that leading psychiatrists who
have worked on it have been inuenced by fund-
ing relationships with the drug industry,
16
but the
APA has taken it in its stride. While such specu-
lation was bound to occur, said David Kupfer,
professor of psychiatry at the University of Pitts-
burgh and chair of the DSM-5 task force, in a press
release in January, it is important to stay focused
on the fact that APA has gone to great lengths to
ensure that DSM-5 and APAs clinical practice
guidelines are free from bias.
Steps taken, Kupfer told Medscape in January,
included limiting task force members to annual
income from industry sources of no more than
$10 000 each and to holding shares in pharma-
ceutical companies worth less than $50 000
limits more stringent than requirements
for staff at the National Institutes of Health,
members of advisory committees for the Food
and Drug Administration, and most academic
d epartments.
17
But corruption-conspiracy theories miss the
point, says Greenberg, who sees instead a con-
federacy of good intentions at
work. The DSM is created by com-
mittees, which is one of the reasons
its such an unwieldy document, and
the committees are made up of experts
in the held, who tend to be people who are
valued and pursued by the drug companies to
do their research.
I dont feel theres a huge conspiracyits
not like the drug companies say to a psychia-
trist, Look, we could really use this disorder in
the DSM, heres hhy grand. They dont have to,
because youve got an entire profession that intel-
lectually is already predisposed to seeing mental
problems as problems that should be treated with
drugs.
Nevertheless, it hasnt helped the cause of
the DSM, or US psychiatry in general, that since
2008 Senator Charles Grassley, the chair of the
US Senate Committee on Finance, has doggedly
unearthed a series of cases in which leading aca-
demic psychiatrists have failed to reveal large pay-
ments from drug companies.
One of the most high prohle cases was that of
Charles Nemero, who in 2008 resigned as chair
of the psychiatry department at Emory University,
Atlanta, aher it was revealed that he had failed to
report more than $1.2m of payments from Glaxo-
SmithKline, despite having signed an undertak-
ing to limit payments to $10 000 a year and while
acting as lead investigator on a National Insti-
tutes of Health study of the companys drugs for
depression.
18-21
In 2009 Nemero was appointed
chair of psychiatry at the University of Miami.
Challenge from biology
A bigger blow to the future of DSM, however, was
landed last month by the US National Institute of
Mental Health. It has launched a stinging attack
on DSM, criticising its lack of validity and
announcing it is re-orienting its research [fund-
ing] away from DSM categories and towards
the establishment of a new classihcation system
based on the biology as well as the symptoms of
mental disorders.
The Research Domain Criteria (RDoC) project,
said Thomas Insel, director of the NIMH, last
month, was nothing less than a plan to trans-
form clinical practice by bringing a new genera-
tion of research to inform how we diagnose and
treat mental disorders.
The long term future of mental health, he
said, lay in the detection of biomarkers: Unlike
our definitions of ischaemic heart disease,
lymphoma, or AIDS, the DSM
diagnoses are based on a
consensus about clusters
of clinical symptoms, not
any objective laboratory
measure.
22
DSM-5: Too many labels?
20 BMJ | 25 MAY 2013 | VOLUME 346
PSYCHIATRY
extensive reviews of the scientihc literature, origi-
nal research, and nationwide held trials.
Those held trials, of course, have caused their
own share of controversy. For Farley and others,
all they showed was that the science [behind
DSM-5] isnt good enough.
Between November 2010 and November
2012 the DSM-5 criteria for 23 disorders were
tested by hundreds of psychiatrists working with
3500 patients across the country. Six diagnoses,
according to the American Psychiatric Associa-
tion, had relatively low reliability scores, which
the report characterised as questionable but
acceptable; among them were two of the DSMs
most common diagnoses: major depressive disor-
der and generalised anxiety disorder.
Legislative threat
All the debates over what should and should not
go in DSM-5 are now academic. But the future of
the manual may face a longer term threat in the
form of the mandatory shih in October 2014 from
ICD-9 to ICD-10 code sets for all healthcare pro-
viders covered by the US Health Insurance Port-
ability and Accountability Act (HIPAA), which
embraces the Medicare national insurance pro-
gramme and Medicaid, the means tested health
programme.
As with ICD-9, the US has modihed ICD-10 for
use in its own medical system, but the dierence
with ICD-10-CM is that it will pull many psychia-
trists and psychologists away from the orbit of
DSM for the hrst time.
Currently, many psychologists utilise the
DSM-IV-TR when diagnosing patients, and the
corresponding DSM-IV-TR codes to submit a
health insurance claim, noted the psychiatric
association in a practice update to members in
February. That was hne under ICD-9 because its
diagnostic codes aligned with those of DSM, but
the ICD-10 codes will be dierent and anyone
who bills DSM codes instead of ICD-10-CM codes
presumably will experience claim denials.
26
Kapur expects a well-rounded discussion at
the Kings conference next month.
It is, he says, easy to criticise, but I think many
people are really ahistorical, because they can
hnd all the faults with DSM but are totally naive
about the total mess that the diagnosis of psy-
chiatric disorders used to be about 40 years ago
. . . What you called schizophrenia in New York
wasnt schizophrenia in London, and we have
made a huge transition since then.
Jonathan Gornall freelance journalist, Suffolk, UK
jgornall@mac.com
peer reviewed.
Starting virtually from scratch, RDoC is for
now a distant challenge to the supremacy of
DSM. It is the way forward, but it is nowhere
near where one can use it to create a diagnostic
system, says Shitij Kapur, dean of the Institute
of Psychiatry at Kings College London, who last
year published a paper with Insel examining why
it has taken so long for biological psychiatry to
develop clinical tests.
23
For now, he says, we have a job to do and
[ICD and DSM] are the tools for today. The dif-
ferences between them are in 95% of the cases
either trivial or only for the connoisseurs, and
the real question now is does the world need a
DSM and an ICD?
In the meantime, he thinks the fuss surround-
ing the new DSM will all go away, because
DSM-5 is not a revolution; it is a mere sensible
incremental improvement. There are people who
are disappointed, because 10 years ago there was
hope that we would have clear biological tests
and that this DSM would radically revise all the
d ehciencies we have in psychiatric disorders.
He says it is not too simplistic to characterise
the controversy over DSM as psychiatry versus
psychotherapy, or medication versus talk ther-
apy, and that those camps are duking out those
tensions against the backdrop of the DSM launch.
This is not to say that those tensions are not real,
or unimportant; its just that they didnt happen
yesterday and DSM-5 doesnt particularly make
them dramatically better or worse.
DSMs influence in UK
On 4 June the Institute of Psychiatry at Kings
will host a two day international conference on
DSM-5 and the future of psychiatric diagnosis.
Although the UK uses the ICD, it would, says
Kapur, be a mistake to assume that DSM has no
eect on British psychiatry.
It would be unwise for the rest of the world
to ignore things that are happening in the US
because it is undoubtedly the
academic and professional pow-
erhouse in psychiatry, generating
so many concepts, ideas, new hnd-
ings, and more papers than any
other country almost by a factor of
10, he says.
When it comes to scientihc dis-
course, he adds, the DSM trumps
the ICD: a British psychiatrist treat-
ing a patient who has schizophrenia is likely to be
reading papers on the subject in which patients
have been diagnosed by doctors using DSM.
As editor of the British Journal of Psychiatry,
Tyrer says he repeatedly reminds American
authors that the UK is one of the last DSM-free
zones in the world, because we havent been inu-
enced by the heavy selling of the DSM.
Nevertheless, if someone says, This person
has a DSM diagnosis of major depression, that
carries much more weight than saying, This
person has a ICD diagnosis of mild depressive
illness.
This was shown by a high prohle British trial
in 2005, in which a man accused of killing his
parents and embarking on a spending spree with
their credit cards avoided a murder conviction by
admitting manslaughter with diminished respon-
sibility on the ground that he had narcissistic
personality disorder.
24
The disorder appears only
in DSM: Narcissistic has never been in the ICD
classihcation and I am conhdent it never will be,
says Tyrer. It is a diagnosis of vanity for both
patients and their doctors.
DSMs defence
One man who will be attending the Kings DSM-5
conference next month is David Kupfer, chair of
the DSM-5 task force, who clearly believes the UK
merits a visit during what for him must surely be a
frenetic post-publication periodand who gives
not the slightest impression of being a man in a
diabolically hot seat.
One of the key messages he is keen to get across
is that DSM-5 remains the best science available
[and] the best manual for helping clinicians care
for patientsand, unsurprisingly, he does not
accept the notion that DSM should step aside in
favour of ICD.
DSM and ICD can be thought of as companion
publications, he says. They are cross-linked so
that a clinician using DSM can use the ICD diag-
nostic coding system required by most of the
worlds health systems. Its important to note
that ICD does not include descriptive diagnostic
criteria, only a listing of disorders. DSM-5 is the
best possible clinical guidebook for the diagnosis
of mental disorders.
Likewise, despite the vehemence of the
National Institute of Mental Healths attack on
DSM, he insists that DSM-5 and
the National Institutes RDoC
represent complementary, not
competing, frameworks. Once the
RDoC eort starts to take shape,
any information or hndings stem-
ming from its research agenda will
be integrated into future DSM edi-
tions to further strengthen patient
care.
He also sidesteps a question about the dim-
culty of having his predecessor as chair of DSM-IV
as one of the hercest critics of DSM-5.
While criticism is bound to occur, he says,
we think it is important to stay focused on the
fact that DSM-5 was developed over the course of
more than a decade with the input of more than
1500 of the best experts in the held and through
I dont feel theres a
huge conspiracyits
not like the drug
companies say to a
psychiatrist, Look, we
could really use this
disorder in the DSM,
heres fifty grand
BMJ | 25 MAY 2013 | VOLUME 346 21
ANALYSIS
A NICE example?
Variation in the
provision of
bariatric surgery
Demand for surgery to treat morbid obesity
outstrips supply. Amanda Owen-Smith and
colleagues find regional commissioning
policies are not consistent with NICE guidance
and provision of surgery varies widely
sumed within a broader clinical guideline on
obesity,
5
which eectively removed the mandate
on PCTs to provide access to surgery, although it
was still recommended as the most cost eective
treatment; the new guideline added that surgery
should be considered as hrst line treatment for
people with a body mass index >50 (box 1).
Access to surgery
Since NICE published the guideline, evidence
has increased that surgery has a positive eect on
weight loss and the resolution of obesity related
comorbidities (box 2, bmj.com).
6

7
Indeed, it is
the only treatment currently available for morbid
obesity that has a reasonable chance of sustained
weight loss.
8

9
The number of NHS operations
has soared, with 8087 weight loss operations
(including revisions) undertaken in 2010-11,
a rise of over 300% since 2006-07.
10
However,
NHS provision is still dwarfed by the rising tide
of morbid obesity (body mass index >40) in
England, which was estimated at 2.7% in 2010,
equivalent to 1.35 million people.
11
Of course, not
everyone who is morbidly obese would choose,
T
he National Institute of Health and Care
Excellence (NICE) has been the most
ambitious attempt to set healthcare pri-
orities more systematically and explic-
itly since the inception of the National
Health Service in 1948.
1
NICE was established
in 1999 with the dual aims of reducing regional
inequality in access to treatments and making
decision making more systematic through explicit
consideration of the clinical and cost eectiveness
of (mostly) new treatments. Despite controversies
related to the time taken to issue recommenda-
tions and the defensibility of its cost eectiveness
thresholds, the NICE decision making process is
respected within the UK and internationally.
2

3

Nevertheless, NICE guidance has met with a
mixed reaction from NHS commissioners, who
have to implement the recommendations, some
of which are mandatory within three months.
Frustrations have arisen over the lack of consid-
eration given to the aordability of implementa-
tion. (NICE recommendations have to be funded
from existing allocations, implying disinvest-
ment elsewhere.) Drawing on the example of
bariatric surgery for morbidly obese patients,
we review the implementation of NICE guidance
and regional specialist commissioning guidelines
over the past 10 years. We examine how they have
aected provision of care and consider what chal-
lenges remain for the new clinical commissioning
groups and NHS England as they take on their full
commissioning roles.
NICE guidance
NICE issued mandatory guidance on bariatric
surgery in 2002.
4
The technology appraisal
instructed primary care trusts (PCTs) to provide
surgery for morbidly obese patients (body mass
index 40 or 35 in the presence of specified
comorbidities) when other interventions for
weight loss had failed. Although PCTs were given
some exibility on the three month implementa-
tion schedule because of shortages in capacity,
the message from NICE was clear: surgery was
both eective and cost eective and should be
made available for patients meeting the crite-
ria when both they and their clinicians think it
is appropriate. In 2006, the appraisal was sub-
Box | NICE criteria for bariatric surgery
Bariatric surgery is recommended if all of the following criteria are fulfilled:

Body mass index or - if other serious disease (eg, type diabetes or high blood pressure)
could be improved by weight loss
All appropriate non-surgical measures have been tried and failed to achieve or maintain adequate
clinically beneficial weight loss for at least months;
Patient has been receiving or will receive intensive management in a specialist obesity service
Patient is generally fit for anaesthesia and surgery
Patient commits to long term follow-up.
Bariatric surgery is also recommended as a first line option (instead of lifestyle interventions or
drug treatment) for adults with body mass index > in whom surgical intervention is considered
appropriate
SUMMARY POINTS
NHS provision of bariatric surgery has risen by
300% over the past six years but less than 0.6%
of those potentially eligible receive treatment
Access to surgery varies widely between regions
and primary care trusts and bears little relation
to stated policy
Access to care risks becoming even more
inequitable as clinical commissioning groups
take on NHS commissioning in an era of
mounting austerity
Better and more transparent approaches to local
priority setting are needed
L
I
F
E

I
N

V
I
E
W
/
S
P
L
bmj.com
Obese patients get inadequate care before and after bariatric surgery,
nds review (BMJ 2012;345:e6890)
Feature: Slimmed down surgery (BMJ 2010;341:c5499)
Clinical review: Surgery for obesity in adulthood (BMJ 2009;339:b3402)
22 BMJ | 25 MAY 2013 | VOLUME 346
ANALYSIS
or be suitable for, surgery because of the lifestyle
restrictions it imposes or the severity of comor-
bidities.
5
Additionally, the hrst national audit of
bariatric surgery suggests that about 31% of bari-
atric operations are currently done in the private
sector, 86% of which are on patients with a body
mass index over 40, implying eligibility for NHS
treatment.
12
Nevertheless, the fact that less than
0.6% of those who are potentially eligible for bari-
atric surgery receive treatment on the NHS, cou-
pled with media reports and qualitative research
documenting dimculties in accessing care, indi-
cates a shortfall in provision or access.
13

14
Data for 2010-11 show substantial regional
variations remain in access to NHS bariatric sur-
gery, with rates ranging from 6/100 000 popula-
tion in the North West region to 32/100 000 in
the East Midlands (table 1). There are no data
available on the prevalence of morbid obesity
by region, and although data from the Health
Survey for England suggest some regional dier-
ences in the prevalence of simple obesity (body
mass index >30; 22-26.5%) this is not sumcient
to explain the variation observed.
10
In addition,
inequalities in provision have widened over the
past hve years, suggesting dierences in imple-
mentation of NICE guidance.
15
Role of regional commissioning policies
As some of the dierences in provision might
be explained by regional variations in special-
ist commissioning policies, we systematically
compared these policies with the NICE criteria
(table 2). When possible, we accessed up to date
documents online. If none could be located we
contacted relevant bodies until they provided the
policies or conhrmed that none were available.
Of the eight sets of specialist commissioning poli-
cies we were able to access, none provided for all
of the patient groups that NICE recommends are
suitable for bariatric surgery. Only half of regions
recommend treatment for those with a body mass
index of 35-39 in the presence of comorbidity,
and only one (Yorkshire and Humber) complies
with the NICE recommendation that bariatric sur-
gery should be hrst line treatment for those with a
body mass index >50.
However, variations in regional policies do not
necessarily reect rates of provision of bariatric
surgery. Indeed, the highest providing region, the
East Midlands, has some of the strictest criteria
for access to surgery, and the North West, which
is the lowest provider, has among the most gener-
ous policies. Although hve of the 10 regions are
ranked roughly as might be expected given varia-
tions in access criteria, levels of NHS provision are
so low throughout England that it seems unlikely
that even people fulhlling the strictest criteria
would easily be able to access NHS treatment.
Our analysis suggests that regional policies for
accessing bariatric surgery are not followed con-
sistently, leading to dierences in local provision.
Other factors that might contribute to the varia-
tion include regional dierences in prevalence
of morbid obesity, the capacity of surgical units,
dierential local funding pressures and priority
setting processes, and dierences in the attitudes
of commissioners, clinicians, and patients to the
acceptability and eectiveness of a surgical solu-
tion to the problem of morbid obesity.
Variation at the primary care trust level
The NHS Atlas of Variation data also show con-
siderable variation in provision by individual
PCTs within regions, with a 93-fold dierential
observed between the highest and lowest provid-
ing trusts.
10
These hndings are consistent with
patient information websites that list the diering
access criteria that people have encountered,
16

and with qualitative research that has found
that some PCTs are adding additional access
criteria to help manage the gap between supply
and demand.
14
For example, PCTs in one area
provided surgery only for those with diabetes,
even though it was openly acknowledged that
this leh other needy patients without access to
care. One consultant in the area said, Ive got
a desk drawer full of patients who arent eligible
[because they arent diabetic] even though theyre
severely depressed to the extent that theyve tried
to commit suicide . . . or theyve got sleep apnoea,
or other signihcant comorbidities.
17
Equitable access
Our analysis raises several questions for patients
and providers seeking to improve access to baria-
tric surgery, and for the new clinical commission-
ing groups and NHS England. The data show
considerable regional variation in the provision
of morbid obesity surgery, and local commission-
ing guidelines suggest that individual access to
treatment remains subject to a postcode lottery.
Regional variation in the availability of care is
widely known to be disliked by patients and cli-
nicians, and was one of the problems that NICE
guidance was designed to overcome.
1
In addition,
overall levels of provision clearly fall far short of
the number of operations that would be needed if
NICE access criteria were to be fully implemented,
implying that some people who could be eec-
tively and cost eectively treated are leh without
access to care.
In this era of austerity, it is unlikely that the
investment needed to fund surgery for morbid
obesity to the level recommended by NICE will
be forthcoming. It will thus fall to more local
decision makers to prioritise patients for treat-
ment and decide what (if any) investment should
be awarded to morbid obesity surgery amid the
many other calls on constrained budgets. For at
least the next two years, clinical commissioning
groups can be expected to be supported by NHS
England, which will issue guidelines on access to
specialised treatments and promote regional con-
sistency in access to care.
18
However, our analysis
suggests that simply issuing further sets of guide-
lines is unlikely to promote either regional equity
in access to bariatric surgery or even access for
most patients eligible for treatment.
Clinical commissioning groups are likely to
hnd it dimcult to eectively prioritise patients for
care in the short term and determine the capacity
Table | Compliance of regional commissioning policies for bariatric surgery with NICE guidance
Region
NICE eligibility criteria
BMI - in presence of
serious comorbidity
BMI - when non-surgical
methods have failed
BMI > as first line
treatment
East Midlands
w1
No No* No
London
wl
No regional guidance
North East
w!
Yes Yes No
Yorkshire and Humber
w/
Yes/No Yes /No Yes
South East Coast
w'
Yes Yes No
West Midlands
w6
No regional guidance
South West
w7
Yes Yes No
East of England
wS
No No No
South Central
w9
No No Unclear
North West
w1u
Yes Yes No
BMI=body mass index. *Must have comorbidity and BMI /'.
Half of PCTs in the Yorkshire and Humber area were operating to NICE criteria and half stipulated patients must have BMI /' and
serious comorbidity to be eligible. Must have type l diabetes or severe sleep apnoea.
Table | Number of bariatric procedures per
population by strategic health authority
area, -*
Strategic health authority Male Female Total
East Midlands 1l '1 !l
London 9 /u l/
North East 9 !! l1
Yorkshire and the Humber S l! 16
South East Coast S l7 1S
West Midlands 6 17 1l
South West ' 1l 1S
East of England 6 11 9
South Central ' 1/ 9
North West ! 1u 6
*Finished consultant episodes with a primary diagnosis of obesity
and a main or secondary procedure of bariatric surgery (Hospital
Episode Statistics).
BMJ | 25 MAY 2013 | VOLUME 346 23
ANALYSIS
required in the longer term. Decisions might be
aided by using the widely supported accountabil-
ity for reasonableness ethical framework, which
argues that any prioritisation structure must meet
the conditions of public accessibility, relevance
to the local population, enforceability, and the
availability of an appeals process.
19
Tools such as
programme budgeting and marginal analysis are
also available to help commissioners make more
informed decisions about the costs and benehts
associated with investing or disinvesting in dier-
ent treatments.
20
This analysis is limited by our inability to accu-
rately estimate the number of people who would
be eligible for treatment under NICE guidance
and the uncertainty surrounding the inuence of
other factors on regional variations in treatment
provision, such as local capacity shortages or
dierent clinician and patient attitudes towards
the acceptability of care. Nevertheless, given that
there are many other examples of treatment areas
where NICE guidance does not seem to be fully
implemented, including multiple sclerosis,
21

diabetes treatment,
22
treatment for sub-fertility,
23

and numerous drug interventions,
24
it is clear
that clinical commissioning groups will struggle
to achieve fairness and equity of provision even
if they are given more sophisticated support to
implement new and existing guidelines. This
requires development of good practice in local
prioritisation using recognised tools and frame-
works so that there is more transparency around
what is being invested in and why, together with
research to better understand need and demand.
Amanda Owen-Smith research fellow, School of Social
and Community Medicine, University of Bristol, Bristol
BSS lPS, UK
Ruth Kipping research fellow and consultant in public
health, School of Social and Community Medicine,
University of Bristol, Bristol BSS lPS, UK and NHS North
Somerset, Public Health, Clevedon, UK
Jenny Donovan professor of social medicine, School of
Social and Community Medicine, University of Bristol,
Bristol BSS lPS, UK
Christine Hine consultant in public health, NHS Bristol
and NHS South Gloucestershire Public Health Directorate,
Emersons Green, Gloucestershire, UK
Christina Maslen clinical effectiveness research lead,
NHS Bristol, Bristol, UK
Joanna Coast professor of health economics, Health
Economics Unit, School of Health and Population
Sciences, University of Birmingham, Birmingham, UK
Correspondence to: A Owen-Smith a.owen-smith@bris.ac.uk
Accepted: 1l March lu1!
Contributors and sources: AOS, JC, and JD have all been involved
in primary qualitative research about the implementation of
NICE guidance in this area, and RK, CH, and CM have all been
involved with the implementation of NICE guidance at the PCT
level. All authors contributed to the idea for this article. The initial
review of regional policies was undertaken by CM and updated
by AOS. AOS undertook the bulk of the literature review, assisted
by JC, JD and CH. Quantitative data were largely researched
and contributed by RK. AOS wrote the article, which has been
reviewed and commented on by all authors. AOS is the guarantor.
Competing interests: This study was undertaken as part of AOSs
NIHR postdoctoral fellowship, but the NIHR had no involvement
in the work or the decision to submit the paper for publication.
Provenance and peer review: Not commissioned; externally
peer reviewed.
1 Oliver A, Mossialos E, Robinson R. Health technology
assessment and its influence on health-care priority setting.
Int J Health Technol Assess luu/;lu:1-1u.
l Hawkes N. NICE goes global. BMJ luu9;!!S:l66-7.
! WHO. Health technology and health policymaking in Europe.
WHO, luuS.
/ NICE. Guidance on the use of surgery to aid weight reduction
for people with morbid obesity: technology appraisal
guidance No /6. NICE, luul.
' NICE. Obesity: the prevention, identification, assessment
and management of overweight and obesity in adults and
children. NICE, luu6.
6 Picot J, Jones J, Colquitt JL, Gospodarevskaya E, Loveman
E, Baxter L, et al. The clinical effectiveness and cost-
effectiveness of bariatric (weight loss) surgery for obesity: a
systematic review and economic evaluation. Health Technol
Assess luu9;1!:1-19u.
7 Schauer PR, Kashyap SR, Wolski K, Brethauer SA, Kirwan JP,
Pothier CE, et al. Bariatric surgery versus intensive medical
therapy in obese patients with diabetes. N Engl J Med
lu1l;!66:1'67-76.
S Colquitt JL, Clegg A, Loveman E, Royle P, Sidhu MK.
Surgery for morbid obesity . Cochrane Database Syst Rev
luu';/:CDuu!6/1.
9 Sjostrom L. Bariatric surgery and reduction in morbidity
and mortality: experiences from the SOS study . Int J Obesity
luuS;!l:S9!-7.
1u NHS Information Centre. Statistics on obesity, physical
activity and diet: England, lu1l. www.ic.nhs.uk/webfiles/
publications/uu!_Health_Lifestyles/OPAD1l/Statistics_
on_Obesity_Physical_Activity_and_Diet_England_lu1l.
pdf.
11 NHS National Obesity Observatory. Prevalence of morbid
obesity among adults aged 16+ years (based on data from
the Health Survey for England). lu11. www.noo.org.uk/
NOO_about_obesity/morbid_obesity/ukprev.
1l Welbourn R, Fiennes A, NBSR Data Committee. The United
Kingdom national bariatric surgery registry. First registry
report to March lu1u. Dendrite Clinical Systems, lu11.
1! Triggle N. Obesity surgery court battle lost by Tom
Condliff. BBC News lu11 Jul l7. www.bbc.co.uk/news/
health-1/!uS97'.
1/ Owen-Smith A, Coast J, Donovan J. I can see where theyre
coming from, but when youre on the end of it, you just want
to get the money & the drug: Explaining reactions to explicit
healthcare rationing. Soc Sci Med luu9;6S:19!'-/l.
1' NHS Information Centre. The NHS atlas of variation in
healthcare. lu1u. www.rightcare.nhs.uk/atlas/qipp_
nhsAtlas-LOW_l6111uc.pdf.
16 British Obesity Surgery Patient Association. PCT surgery
funding criteria. lu1l www.bospa.org.
17 Owen-Smith A. Knowing about healthcare rationing. PhD
thesis. University of Bristol, luuS.
1S NHS Commissioning Board. Securing equity and excellence
in commissioning specialised services. lu1l. www.
commissioningboard.nhs.uk/files/lu1l/11/op-model.pdf.
19 Daniels N, Sabin J. Accountability for reasonableness: an
update. BMJ luuS;!!7:a1S'u.
lu Ruta D, Mitton C, Bate A, Donaldson C. Programme
budgeting and marginal analysis: bridging the divide
between doctors and managers. BMJ luu';!!u:1'u1.
l1 Hitchen L. NICE recommendations have had little
effect on multiple sclerosis services five years on. BMJ
luuS;!!7:a7!/.
ll Kmietowicz Z. Nearly two thirds of people with diabetes in
England dont receive all recommended care, audit finds.
BMJ luu9;!!9:bl6S'.
l! Hughes D. MPs condemn regional variations in IVF
provision. BBC News lu11 Jun 7. www.bbc.co.uk/news/
health-1!67u61'.
l/ Health and Social Care Information Centre. Use of NICE-
appraised medicines in the NHS in England, in lu1u and
lu111. www.ic.nhs.uk/pubs/niceappmed1u11.
ANSWERS TO ENDGAMES, p 40 For long answers go to the Education channel on bmj.com
CASE REPORT A treatable cause of shortness of breath
This patient has chronic thromboembolic pulmonary hypertension (CTEPH).
Investigations needed to confirm the diagnosis and elucidate the cause of
pulmonary hypertension comprise chest radiography, electrocardiography, full
pulmonary function testing, echocardiography, ventilation-perfusion scanning,
computed tomography angiography, and right heart catheterisation.
Patients with suspected CTEPH should be referred to the regional specialist
service for confirmation of the diagnosis and appropriate specialist treatment.
All patients with a diagnosis of CTEPH should be referred to the nearest
specialist centre for consideration of pulmonary endarterectomy.
STATISTICAL QUESTION
P values or confidence intervals?
Statements c and d are true, whereas a and b are false.
PICTURE QUIZ An unusual burn
The photograph shows a full thickness burn to the volar aspect
of the wrist, total body surface area about .%.
Initial management is to assess and resuscitate using
advanced trauma life support and emergency management of
severe burns principles, remove the burning source, and cool
with tepid running water. Further assessment of burns includes
full history (to identify the source and mechanism), careful
examination, appropriate investigations, discussion with local
burns unit, and application of non-adherent dressing.
In addition to cutaneous injury at the contact points,
complications specific to electrical burns include cardiac
arrhythmias, rhabdomyolysis from thermal damage to muscle,
acute renal failure from myoglobinuria, hepatic dysfunction,
acute fracture, compartment syndrome, and cataracts.
24 BMJ | 25 MAY 2013 | VOLUME 346

LETTERS
Letters are selected from rapid responses posted on bmj.com. Aher editing, all letters are
published online (www.bmj.com/archive/sevendays) and about half are published in print
To submit a rapid response go to any article on bmj.com and click respond to this article
novo lipogenesis is low during carbohydrate
overfeeding, a common misconception. As
the last reference Cottrell cites says, body fat
stores can be increased not by conversion of
the carbohydrate to fat. Instead, the oxidation
of dietary fat [is] suppressed and fat storage
thereby increased.
The past years has shown us that we need
new information and better science because
the conventional wisdom isnt working. People
have a spectrum of metabolic disorders; it is
incumbent on us to challenge conventional
thinking when it is clearly failing.
Gary Taubes co-founder , Nutrition Science Initiative,
San Diego, California, USA gataubes@gmail.com
Full response at www.bmj.com/content//bmj.f/
rr/ .
1 International Sugar Organisation. Quarterly Market Outlook.
lu11. www.sugaronline.com/Samples/ISO_QMO.pdf .
l Matthews DR, Levy JC. Impending type l diabetes. Lancet
luu9 ; !7! : l173 -9.

Am I missing something?
Taubes contrasts two obesity hypotheses
excess energy intake versus a hormonal
disorder.
It becomes clear, however, that

the second theory is not about a disorder, but
about excess consumption of carbohydrates
causing obesity, with insulin implicated as
the mechanism. A disorder normally means
pathology of a particular system, or a
derangement or abnormality of function; a
morbid physical or mental state according to
an online medical dictionary. This means there
is no disorder here, just normal biochemistry. So
why the apparent insistence on the disordered
nature of this theory?
The two theories seem to be excess
consumption of calories versus excess
consumption of carbohydrates. Given that
a large part of calorie intake is made up of
carbohydrates (recommended to be -%),

these theories arent worlds apart, so why are
they presented as being so?
Furthermore, given that both theories concern
overconsumption, presumably they involve
the same causative mechanism with regard to
appetite? In that case, the presumption that
pathological appetite mechanisms underlie
this disease seems odd. Human levels of
consumption have as much to do with availability
of food as with physiological drivers of appetite. I
for one dont need to be hungry to eat when food
THE SCIENCE OF OBESITY
Essays premises are incorrect
In attempting to resurrect a long discarded
hormonal theory of obesity, Taubes falls into a
circular argument based on incorrect premises
and an unproved assumption.

The circularity rests in his claim that
overconsumption of carbohydrate
causes a hormonal response that leads
to overconsumption of food energy, with
overconsumption being a consequence of the
hormonal response rather than its cause.
An axial, but incorrect, premise is that
increased consumption of food energy as
carbohydrate, especially sugars, has been
the dominant ecological association with
the increase in obesity worldwide. Food and
Agriculture Organization data show that the
supply of sugars for human consumption has
remained static, per head of population, for the
past years. In contrast, food energy supply
from other sources has increased.

A second incorrect assertion is that
consumption of carbohydrates leads to
chronically raised insulin concentrations, which
favour fat storage. In normal people, insulin
remains low and stable for many
years,
and carbohydrates are not

uniquely obesogenic.

A third incorrect assertion is
that obesity can be attributed to
the conversion of carbohydrate
to fat. This is an unsatisfactory
explanation of obesity, because
this route is a minor pathway
to depot fat in humans, even
under conditions of substantial
overfeeding of sugars to obese
subjects.

An unproved assumption is that the
hypothetical diversion of carbohydrate energy
into fat storage leaves the subject hungry, thus
stimulating overeating.
Taubes dismisses the energy balance
hypothesis of obesity on the grounds that it
has not led to a universally successful form
of treatment, but he offers only anecdotal
evidence for his alternative view. He also offers
no explanation of why disordered hormonal
regulation affects only some people.
Richard Conrad Cottrell director general , World
Sugar Research Organisation, London SW1V !LX, UK
rcottrell@wsro.org
Competing interests: RCC is an employee of the World Sugar
Research Organisation, a not for proht scientihc research
organisation funded by the sugar industry.
1 Taubes G. The science of obesity: what do we really know
about what makes us fat? An essay by Gary Taubes. BMJ
lu1! ; !/6 : f1u'u . (16 April.)
l World Sugar Research Organisation. WSRO Report on Trends
in Per Capita Sugar Supply, 1961-luu7. www.wsro.org/
Portals/1l/Docs/report-on-trends-in-per-capita-sugar-supply-
1961-to-luu7-world-and-regional-level.pdf .
! Tabak AG, Jokela M, Akbaraly TN, Brunner EJ, Kivimki M, Witte
DR. Trajectories of glycaemia, insulin sensitivity, and insulin
secretion before diagnosis of type l diabetes: an analysis from
the Whitehall II study. Lancet luu9 ; !7! : ll1' -l1.
/ Food and Nutrition Board, Institute of Medicine, National
Academy of Sciences. Dietary reference intakes for energy,
carbohydrates, fiber, fat, fatty acids, cholesterol, protein and
amino acids (macronutrients). National Academic Press, luu'.
' Hellerstein MK. No common energy currency: de novo
lipogenesis as the road less travelled. Am J Clin Nutr
luu1 ; 7/ : 7u7 -3.

Authors reply
Cottrell cites a World Sugar Research
Organisation report claiming that world
consumption of caloric sweeteners (such as
sucrose and high fructose corn syrup) has been
static for decades. These data are difficult to
reconcile with International Sugar Organisation
reports that world sugar consumption rose
between and by an
average of .% yearly, nearly
twice the population growth
rate.

Such trends can neither
confirm nor refute the hormonal-
regulatory defect hypothesis or
sugars role in obesity. Cottrell
assumes incorrectly that caloric
sweetener consumption must
increase to explain the obesity
epidemic. If a threshold effect
is at work, then prevalence can
increase if sugar consumption is above that
threshold. This is another argument for well
controlled experiments to establish causality.
Cottrell claims that insulin concentrations
remain low in healthy people, but the study he
cites reports a long term deterioration of insulin
sensitivity in this population. The accompanying
editorial suggests that it is the effect of
progressive inactivity and middle age spread.
It could also be the effect of a dietary trigger of
progressive inactivity and middle age spread.

Such observations cant differentiate.
Cottrell suggests that the hormonal-regulatory
defect hypothesis is incorrect because de
BMJ | 25 MAY 2013 | VOLUME 346 25
LETTERS
is available. Or am I missing something again?
It seems worrying that someone who is
already convinced that obesity is due to an
unidentified disordered environmental stimulus
causing excess carbohydrate intake, and in turn
causing a disordered hormonal response, is
leading an organisation with vast resources to
go looking for these mechanisms. And this is
despite perfectly adequate explanations being
explicit and implicit in his article.
Please tell me Im missing something.
Ben Bradley general practitioner, Meuchedet Health
Care Organisation, Ramat Beit Shemesh, Israel
bdbradley7u@hotmail.com
Competing interests: BB needs to lose a little around the
midril.
lu1!;!/6:f1u'u. (16 April.)
l Institute of Medicine. Dietary reference intakes for energy,
carbohydrate, fiber, fat, fatty acids, cholesterol, protein, and
amino acids. luul. www.iom.edu/Reports/luul/Dietary-
Reference-Intakes-for-Energy-Carbohydrate-Fiber-Fat-Fatty-
Acids-Cholesterol-Protein-and-Amino-Acids.aspx.
Lets act on available evidence
Taubes argues that we start anew in obesity
research.
1
From a public health perspective, I
have five reflections on his essay.
Firstly, he juxtaposes the energy balance
against a failure of hormonal/metabolic
regulation. However, the two are not mutually
exclusive. Did previous generations of
researchers really consider energy imbalance as
the cause of obesity without thinking that other
factors might influence that balance?
Secondly, in saying All we have are
speculations, Taubes is throwing out the baby
with the bathwater. Granted, some research is
substandard, but he ignores the excellent work
that has also been done, while paradoxically
citing some of it.
Thirdly, Taubess views are paralysing. If we
dont know anything, we cant do anything.
We are reduced to waiting for the results from
research funded by the Nutrition Science
Initiative. (But that will be rewarding: according
to the organisations website, it will find
out, once and for all, what we need to eat to
be healthysounds more like miracle than
science.)
Fourthly, Taubes does not mention societal
influences. Our metabolic hard wiring may
make us vulnerable to overdosing with sugar,
but our basic metabolic processes have not
changed over the past decades, whereas
our waistlines have expanded. The question
is what environmental factors drive us to
consume more energy than we use and how
these changed over time.
Finally, Taubess insistence on accepting
evidence only from randomised controlled
trials is too restrictive. This means discarding
promising interventions to reduce the burden
of obesity, such as limits on food advertising
to children
2
and taxation of sugar sweetened
drinks.
3
No trial evidence of such population
targeted interventions is available and possibly
never will be.
New good quality research is welcome. But
rather than wait for perfect evidence, we have
to act on best available evidence.
J Lennert Veerman senior research fellow, School of
Population Health, University of Queensland, Herston,
QLD /uu6, Australia l.veerman@sph.uq.edu.a
Full response at www.bmj.com/content/3/6/bmj.f1u5u/
rr/6/2372.
lu1!;!/6:f1u'u. (16 April.)
l Cairns G, Angus K, Hastings G, Caraher M. Systematic reviews
of the evidence on the nature, extent and effects of food
marketing to children. A retrospective summary. Appetite
lu1!;6l:lu9-1'.
! Brownell KD, Farley T, Willett WC, Popkin BM, Chaloupka
FJ, Thompson JW, et al. The public health and economic
benefits of taxing sugar-sweetened beverages. N Engl J Med
luu9;!61:1'99-6u'.
RESEARCH PAPER OF THE YEAR AWARD
Shortlisted paper does not
include all randomised patients
I was surprised that the Edwards-SAPIEN aortic
valve trial by Makkar et al was shortlisted for
the Research Paper of the Year award 2u13
because a recent BMJ article highlighted how
the investigators excluded 2u% of the patients
randomly allocated into it.
1-3
The shortlisted paper reports on the two
year mortality outcome of Edwards-SAPIEN
valve insertion versus usual medical care
in 358 patients with severe and inoperable
aortic valve stenosis.
2
It is an update of an
earlier paper by the same investigators, and
both papers report the superiority of Edwards-
SAPIEN valve insertion.
2

/
Unfortunately,
neither of the papers discloses the existence
of a randomised continued access protocol,
which randomised a further /1 patients
to Edwards-SAPIEN valve insertion and 5u
patients to usual medical care. The data on
these 91 patients have not been published
but are contained in a limited format within a
briefing document produced for the US Food
and Drug Administration (FDA) in July 2u11.
5
Among these undisclosed participants,
32% (13//1) randomised to valve insertion
died compared with 2u% (1u/5u) randomised
to usual medical care. This equates to one
additional death for every eight patients
receiving the valve. At the time of publication of
the shortlisted paper the Edwards-SAPIEN trial
would have accumulated more than two and
a half years of mortality data; it will have now
accumulated more than three and a half years of
data concerning deaths in these 91 patients.
If patients, doctors, and policy makers are to
make truly informed decisions then full disclosure
and analysis of all patients who were randomised
into the Edwards-SAPIEN valve trial are needed.
Michael A Crilly senior lecturer in clinical
epidemiology, Aberdeen University Medical School,
Aberdeen ABl' lZD, UK mike.crilly@abdn.ac.uk
Competing interests: I have previously advised NHS Grampian
Health Board concerning the provision of trans-aortic valve insertion
(TAVI) and I have also been involved in assessing whether individual
patients with inoperable aortic stenosis should be referred for TAVI
based on the data published by Makkar and colleagues.
1 Groves T. Research Paper of the Year award lu1!. BMJ
lu1!;!/6:fl'1l. (l/ April.)
l Makkar RR, Fontana GP, Jilaihawi H, Kapadia S, Pichard AD,
Douglas PS, et al; PARTNER Trial Investigators. Transcatheter
aortic-valve replacement for inoperable severe aortic
stenosis. N Engl J Med lu1l;!66:1696-7u/.
! Van Brabandt H, Neyt M, Hulstaert F. Transcatheter
aortic valve implantation (TAVI): risky and costly. BMJ
lu1l;!/':e/71u.
/ Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson
LG, et al; PARTNER Trial Investigators. Transcatheter aortic-
valve implantation for aortic stenosis in patients who cannot
undergo surgery. N Engl J Med lu1u;!6!:1'97-6u7.
' FDA. SAPIEN THV briefing documentadvisory committee
meeting. lu11:9'-9.
NO DOCTOR SHOULD BE UNTOUCHABLE
Is the GMC itself untouchable?
My experience of the General Medical Council
is similar to that of Wilmshurst.
1
I forwarded the
GMC several hundred pages of evidence of at
least 2u, very serious, acts of gross professional
misconduct by a senior doctor, the details of
which are extraordinary and shocking.
Not only did the GMC refuse to investigate, but
none of its replies mentioned let alone disputed
any aspect of the evidence or misconduct. The
refusal was point blank. When I re-presented
the complaint on two further occasions over
the years, asking why the complaint had not
been entered into the normal GMC complaints
procedure, the GMC wrote to say that it would
refuse to correspond with me further.
Wilmshurst expressed concerns about
openness, whistleblower protection, and the
current libel laws. My concern is the probity
of the GMC itself. Is the GMC one of the
untouchables?
Jack Gilliat consultant physician, London E1' !HQ,
UK gilliatj@doctors.org.uk
1 Wilmshurst P. No doctor should be untouchable. BMJ
lu1!;!/6:fl!!S. (1S April.)
26 BMJ | 25 MAY 2013 | VOLUME 346
OBSERVATIONS
Are current guideline daily amounts
(GDAs) fit for purpose? With a
worsening obesity crisis and type l
diabetes rising in prevalence, this
question is pertinent. But before
answering it we need to understand
the history of dietary advice.
In luu! the World Health
Organization stated that added
sugars (specifically non-milk
extrinsic sugar) should contribute
no more than 1u% of total energy
intake.
1
This was in line with the
1991 recommendations of the UK
governments Committee on Medical
Aspects of Food and Nutrition Policy.
l

The committee also recommended
that consumption of fruit, vegetables,
potatoes, and bread (containing
intrinsic sugars) should increase by
'u%. Consequently, Rayner et al,
although saying that added sugars
contribute no more than 1u% of total
energy, suggested that total sugars
should contribute to lu% of GDAs.
!
This nutritional advice has formed
the basis of UK food labelling since
luu! and subsequently influenced
European legislation. I believe
that not only has this advice been
manipulated by the food industry for
profit but that added sugar is a risk
factor for obesity and diet related
disease.
In luu9 the American Heart
Association published a scientific
statement in Circulation, Dietary
sugar intake and cardiovascular
health.
/
Acknowledging that the
average US citizen was consuming a
staggering ll teaspoons of added
sugar a day, greatly exceeding
discretionary calorie allowances, the
paper stressed an upper limit of 1uu
kilocalories a day from added sugar
for a woman (six teaspoons) and 1'u
kcal a day for a man (nine teaspoons).
The typical calorie allowance for a /-S
year old child should be a maximum
of three teaspoons a day. Although
a well balanced diet may contain
intrinsic sugars in the form of whole
fruit, vegetables, dairy products,
and many grains, the body does
not require any carbohydrate from
added sugar. Since the American
Heart Association publication, almost
four years ago, several randomised
controlled trials and observational
studies have implicated sugar
consumption with rising rates of
obesity and type l diabetes.
However, the food industry
continues to adopt strategies
to deny sugars role as a major
causative factor in what now
represents the greatest threat to
our health worldwide: diet related
disease. It took 'u years from the
first publication (in the BMJ) linking
smoking to lung cancer before
the introduction of any effective
legislation because Big Tobacco
successfully adopted a strategy of
denial, planting doubt, confusing the
public, and even buying the loyalty of
scientists, all at the cost of millions of
lives. The same corporate playbook
has been adopted by Big Food.
'

A recent European study concluded
that consumption of just one sugar-
sweetened drink a day increased the
risk of type l diabetes by ll%.
6
The
chief spokesperson for the British
Soft Drinks Association downplayed
its results, suggesting that it hadnt
taken into account family history
and, unsurprisingly, that soft drinks
were safe to consume in moderation.
A spokesperson for Diabetes UK
stressed that the findings were not
definitive and suggested that
maintaining a healthy weight was the
most effective way to prevent type
l diabetes. However, the studys
authors had said that the findings
were not necessarily brought on by
obesity but by the high sugar content
of the beverage.
7
A recent longitudinal
cohort study involving 17' countries
showed that, for every additional 1'u
sugar based kilocalories consumed
daily (typical of a can of cola), there
was a massive 11-fold increase in the
risk of developing type l diabetes
independent of body mass index and
physical activity levels.
S
Diabetes UKs corporate partners
include Abbott, the parent company
of Abbott Nutrition, which produces
Isomil Similac, a baby formula that
the paediatric endocrinologist
Robert Lustig has described as the
equivalent of a baby milkshake.
9

(However, a spokesperson for
Diabetes UK has told the BMJ that
the idea that the organisation might
have a conflict of interest because of a
relationship with a corporate partner
was absurd.) Even the UKs most
trusted source of dietary information,
the British Dietetic Association
(BDA), has failed to acknowledge
the adverse effect of excess sugars
on health. I was disturbed to learn
that its food fact sheet states that the
only problem directly linked to sugar
is tooth decay, and the association
flatly denies that eating too much
sugar causes diabetes.
1u
Unlike the
American Dietetic Association
whose sponsors include Coca
Cola, Pepsi-Co, and Kelloggsthe
BDA is not so explicit in revealing
its corporate partnerships. But in
an email exchange with obesity
researcher Zoe Harcombe, the BDAs
partnerships and sponsorship officer,
Jo Lewis, stated that the association
had been delighted to work with the
Sugar Bureau.
11

Another tactic that the food
industry has effectively deployed to
shift the responsibility for obesity
on to the individual is exaggerating
the emphasis on physical activity.
The industry even associates
junk food with sport, allowing the
major food corporations to peddle
pathology with impunity. The recent
London Olympics was dominated by
advertising for junk food and sugary
drinks. And the confectioner Mars is
the official sponsor of the England
football team. A regular sized Mars
bar contains almost triple the amount
of added sugar for an S year old child
that would be recommended by the
US Department of Health and Human
Services dietary guidelines but
represents only !S% of the UK GDA
for total sugars.
Foods that we perceive as junk
are only half the problem. In the
United States a third of added sugar
consumption comes from sugar
sweetened drinks and a sixth comes
from food items such as chocolates,
ice creams, and biscuits, but half
comes from foods that wouldnt
normally be thought of as having
added sugar, such as ketchup,
salad dressings, and bread. Just as
in Europe, US food labels contain
information on total sugars per
serving but do not differentiate
between naturally present and added
sugars. (In the US there is no GDA
for sugar as it is not regarded as a
nutrient.) It is therefore extremely
difficult for consumers to determine
the amount of added sugars in
foods and beverages. One can of
regular cola contains nine teaspoons
of added sugar, which is triple
the luu9 upper limit intake that
the US Department of Agriculture
suggested for an S year old child.
The UK GDA label describes these
nine sugar lumps as !9% of the
guideline daily amount. Based on
this false reassurance, it would be
understandable for parents to believe
it is safe for their child to drink two
and a half cans a day.
Its time for the UKs Scientific
Advisory Committee on Nutrition
and the Department of Health to
act swiftly, as the dietary advice on
added sugar is in desperate need of
emergency surgery.
Aseem Malhotra is interventional
cardiology specialist registrar, Royal Free
Hospital, London
aseem_malhotra@hotmail.com
Provenance and peer review: Commissioned;
not externally peer reviewed.
FROM THE HEART Aseem Malhotra
Dietary advice on added sugar needs emergency surgery
Foods that we think of as junk are only half the problem
One can of regular cola
contains nine teaspoons of
added sugar, triple the upper
limit . . . suggested for an 8
year old child
BMJ | 25 MAY 2013 | VOLUME 346 27
OBSERVATIONS
MEDICINE AND THE MEDIA
Jolie, genes, and the double mastectomy
Angelina Jolies announcement of her risk reducing surgery brought breast cancer to the top of the news agenda. But
her BRCA1 mutation also drew attention to the question of who controls access to genes, writes Richard Hurley
Today it is possible to hnd out through a blood
test whether you are highly susceptible to breast
and ovarian cancer, and then take action, the
Oscar winning Hollywood actor Angelina Jolie
said in the New York Times on 14 May.
1
The
seemingly extreme action that she took was
prophylactic bilateral mastectomy. And the tests
for breast cancer that she mentioned are contro-
versial, because they are controlled by a single
commercial company that has been allowed to
patent the underlying genetic sequences.
Less than three weeks aher her breast recon-
struction the New York Times ran Jolies opinion
article to explain why she had made her deci-
sion. To be proactive and to minimize the risk
as much I could, the star of the Tomb Raider
hlms wrote.
1
In 1000 frank words the 37 year old described
wanting to take control of the inherited
faulty gene BRCA1 that she carries, the same
mutation that led to ovarian cancer in her grand-
mother and her mother, who died aged 56. The
mastectomy had reduced Jolies lifetime risk
of breast cancer from 87% to 5%, she wrote,
although the risk is dierent in the case of each
woman. She has a 50% risk of ovarian cancer
and intends next to undergo an oophorectomy.
Jolie decided to go public because there are
many women who do not know that they might
be living under the shadow of cancer. They do
now: the original New York Times Twitter post
was retweeted almost 5000 times and received
almost 2000 comments, many from women with
BRCA mutations. Her story made the front pages
of most if not all the UK national newspapers.
Now Angelina cant wait to marry Brad sim-
pered the Mail,
2
and a Google search of online
news postings yielded more than 2000 results.
The press have handled it responsibly. This
might alert women with a family history to at
least seek genetic counselling, the surgical
oncologist Michael Baum, who specialises in
breast cancer, told the BMJ. Jolies a great
ambassadorI cant help but admire her, he
said.
The US National Association of Science Writ-
ers blogger Tabitha M Powledge described the
story as the single most-blogged-about medi-
cal topic in the past hve years.
3
The story immediately became the media
pundits must mention topic, and health cam-
paigners, politicians, and celebrities have lined
up to heap admiration on Jolie. She is brave
and an inspiration to many, said the UK for-
eign secretary, William Hague.
4
The television
presenter Sharon Osbourne, who underwent the
same procedure last year, called Jolie a vision.
5
Jolie started the procedure in February, with
surgery to preserve her nipples. She had the
major surgery two weeks later: You wake up
with drain tubes and expanders in your breasts.
It does feel like a scene out of a science-hction
hlm. But days aher surgery you can be back to
a normal life.
1
Meanwhile, with Jolies approval, her doctor
has detailed the surgical regimen. On the web-
site of the Pink Lotus Breast Center, Jolies Bev-
erly Hills clinic, the breast surgeon Kristi Funk
blogged that each womans case was dierent
but that the important thing is to be aware of
your options.
8
She also explained the stages
in identifying risk, considering mastectomy,
and the surgical procedure, along with details
of the drugs and supplements that Jolie took
to prepare for and recover from the operations.
Jolie is also known for her humanitarian con-
science: shes a special envoy of the UN High
Commissioner for Refugees.
Unsurprisingly, then, she used her New York
Times article to draw attention to the dispro-
portionate burden of the cost of paying for
BRCA mutation testing. The cost of testing
for BRCA1 and BRCA2, at more than $3000
[E1980; t2330] in the United States, remains
an obstacle for many women, she wrote.
Part of the reason for this hehy price tag is
that Myriad Genetics, a US biotechnology com-
pany, holds patents for the BRCA1 and BRCA2
sequences and so has a monopoly, at least in
the United States, on the tests for defects in
these genes that cause cancer. Aher publica-
tion of Jolies New York Times article, Myriads
share price closed up 3%.
9
Jolies announcement comes at a key time,
because these patents begin to expire in 2014.
Indeed, on 15 April 2013 the American Civil
Liberties Union and the Public Patent Founda-
tion, on behalf of medical professionals, geneti-
cists, and patients, argued at the US Supreme
Court that these human genes and mutations
were natural and therefore not patentable.
10

But the company has argued that specihc iso-
lated DNA molecules should be eligible for pat-
enting. The court is expected to rule before July.
Jolie described being empowered by dou-
ble mastectomy and has found her reconstructed
breasts beautiful.
But Angelina Jolies a one o, Baum told
the BMJ. Shes exceptional in every way. For
the majority of women it is catastrophic. Its a
threat to their femininity and self conhdence.
Its essential that there is counselling and psy-
chological support.
Regardless, for the worlds most beautiful
woman, as decreed by numerous glossy maga-
zines over the years, to come out smiling aher
such a challenge to her womanhood, coupled
with the unerring support of Brad Pitt, People
magazines one time sexiest man alive, surely
means that the topics of breast and ovarian
cancer and radical surgery are no longer o
limits.
Richard Hurley, deputy magazine editor, BMJ
rhurley@bmj.com
Competing interests: None delcared.
Provenance and peer review: Not commissioned; not peer
reviewed.
Jolies story made the front pages of most if not all the UK national newspapers
28 BMJ | 25 MAY 2013 | VOLUME 346
PERSONAL VIEW
Admission to hospital could be considered a disease
Being in hospital is risky, but Hugh McIntyre considers whether admission should be considered a disease in itself

contemporary medical dictionary
dehnition of disease is an
impairment of the normal state of
the living animal . . . or one of its
parts that interrupts or modihes
the performance of the vital functions . . .
typically manifested by distinguishing signs and
symptoms, and is a response to environmental
factors (as malnutrition, industrial hazards, or
climate), to specihc infective agents (as worms,
bacteria, or viruses), to inherent defects of
the organism . . . or to combinations of these
factors.
1
In Western populations, substantial disease
leads to hospitalisation, and, increasingly,
elderly people have more than one diagnosis.
In such patients, multiple drugs and physical
and cognitive frailness can combine to create
a complex, fragile condition. These people can
be vulnerable to considerable decompensation
when confronted by even small perturbations
such as a fall or mild infection. Admission
to hospital carries its own risks at a time of
enhanced vulnerability. Even when patients
respond to treatment there is a transitional
phase immediately aher hospital, commonly
manifesting as readmission.
Readmission, although considered a
predictable and potentially avoidable hazard, is
common and expensive, occurring in as many
as a hhh of Medicare patients in the United
States for whom data is collected.
2
The rate
in England is around 7%, reecting not only
dierent healthcare practices but also the fact
that Medicare predominantly covers people
older than 65.
3
Risk of readmission is highest in the
hrst few days aher discharge and then falls
exponentially. Aher 30 days the risk is similar to
that of someone who has not been in hospital.
Increasing rates of readmission led Medicare in
2008 to propose including them as a measure
of emciency; this was enacted in law in the
Aordable Care Act in 2012.
4
In 2011-12
the NHS operating framework proposed that
NHS hospitals should not be reimbursed for
readmissions that occur within 30 days.
5
These actions have two consequences.
Firstly, hospital management (and nascent
commissioning structures) is increasingly
focused on readmission. Secondly, it is implied
that readmission is in the hospitals control.
Although some readmissions can be predicted,
we lack evidence for the eectiveness of systems
designed to prevent admissions. This perhaps
reects the fact that readmission is not simply
related to recent admission to hospital nor to the
presenting disease or its severity, and sometimes
readmission is unavoidable.
6
For about a third
of patients the cause of readmission is the same
as that of the preceding admission; for the other
two thirds, diagnoses tend to cluster around
infection, heart and lung disease, metabolic
disturbance, and trauma, including falls.
This may explain why simple evaluations of
the patient, the presenting illness, coexisting
disease, and the number of previous hospital
visits have not helped.
A recent perspective in the New England
Journal of Medicine further challenges
conventional thinking. Harlan Krumholz,
professor of cardiology at Yale School of
Medicine, asked whether when evaluating
readmission disproportionate attention to the
hospitalisations cause . . . may be misdirected.
Patients who were recently hospitalised are not
only recovering from their acute illness but
also experiencing a period of generalised risk
for a range of adverse health events. This is a
condition that may be better characterised as a
post-hospital syndrome, an acquired, transient
period of vulnerability that might derive as
much from the . . . physiological stresses that
patients experience in the hospital as they do
from the lingering eects of the original acute
illness.
7
Beyond the presenting illness, potential
stresses for patients in hospital include not
just the physical impact of sleep deprivation,
undernourishment, and pain, but also the
psychological impact of disrupted personal
routines in the face of unpredictable schedules
and new and changing clinical sta. It is
unsurprising that confusion is common,
especially in older patients.
Even healthy patients mood and immunity
can be aected by sleep and nutritional
disturbance, and such challenges might lead
to unpredictable but substantial impairments
during the early recovery period, an inability to
fend o disease, and susceptibility to mental
error, Krumholz writes. In addition, the
prescribing of new, ohen potent, drugs, perhaps
in combination with existing treatment, can
aect normal physiology and cognition. Simple
deconditioning, which occurs rapidly in older
patients in hospital, will compound the eect of
residual stressors.
Perhaps what makes people vulnerable to
readmission immediately aher hospitalisation
is more likely to be the adverse eect of the
stresses placed upon them during their hospital
stay rather than pre-existing diseases and their
presenting illness. But if it is hospital admission
that leads to readmission, then should we
consider admission to hospital a disease in
itself?
Admission to hospital is certainly an
impairment of the normal state of the living
animal associated with changes to the
performance of the vital functions. Although
residence in a hospital bed cannot be said to be a
distinguishing sign or symptom, readmission
could be thought of as a symptom.
If the concept of post-hospital syndrome
is to be accepted then it reects a response to
environmental factors. Indeed, admission to
hospital could be likened to a disease resulting
from exposure to hazard, in which hostile
environments or agents produce a predictable
pathological response. Such a proposition
could be explored scientihcally, for example, by
the measurement in healthy individuals of the
response of known physiological variables to
similar stressor exposure.
Krumholz concludes with the advice that, if
his construct is valid, then new approaches are
needed to make hospitalisation less toxic.
As a starting point, would framing hospital
admission as a disease help to reorient our
thinking?
Hugh McIntyre is consultant physician, Conquest Hospital,
The Ridge, Hastings, East Sussex TN RD, UK
hugh.mcintyre@esht.nhs.uk
Provenance and peer review: Not commissioned; not
externally peer reviewed.
If it is hospital admission that leads to
readmission, then should we consider
admission to hospital a disease in itself?
A
BMJ | 25 MAY 2013 | VOLUME 346 29
OBITUARIES
Harry Keen
Pioneering diabetes researcher, staunch NHS supporter, and champion of people with chronic disease
Study, which helped to establish the concept of
separate glucose thresholds for microvascular
and arterial disease. He was the natural choice
to chair the 1980 WHO expert committee on
diabetes mellitus, which introduced the concept
of impaired glucose tolerance and laid the basis
for our current dehnition of diabetes.
His experience of an early diabetes clinic
where hve doctors or more sat at desks in the
same large room and exchanged muttered
comments with their patientsled him to
conclude that 10 minutes twice a year was
not enough for the adequate management of
diabetes. He went on to pioneer the diabetes day
centre.
Another idea whose time had come was the
development of the diabetes nurse specialist,
which he regarded as the most important
development in the treatment of diabetes
since the discovery of insulin. He took to heart
Robin Lawrences dictum that the patient with
diabetes must learn to be their own doctor, and
pressed for the involvement of primary care in
diabetes management. He himself worked long
past retirement with the largely South Asian
population attending his sons GP surgery.
Fatalism too ohen characterised diabetes
care in the early part of his career, and Harry
helped the transition to outcome measures
and standards of care. His maxim was to think
globally, act locally. The global component was
achieved with the Declaration of St Vincent,
which set international benchmarks for diabetes
therapy, and the local component laid the
foundation for our national service framework
for diabetes.
It is a measure of the man that his other
landmark achievements need to compete
for mention, but his team was the hrst
to measure microalbuminuria and to
appreciate its signihcance as the harbinger of
diabetic nephropathy, and he also invented
subcutaneous insulin pump therapy.
Harry was mentor to a generation. His
acolytes have sprouted in all directions. What
cannot be penned on paper is the humanity, the
avuncular calm, the resonant probing voice,
the quirky humour, the quizzical smile, and the
burning passion to make things better.
He leaves his wife, Nan; a daughter; and a
son.
Edwin Gale emeritus professor, School of Clinical
Sciences, University of Bristol, Bristol UK
edwin.gale@bristol.ac.uk
Harry Keen was acting as a GP locum in Eltham
when he was called to see a boy whose measles
had been complicated by bronchitis. He
examined the boy, leh a prescription, collected
his two shilling fee, and said he would be back
in a couple of days.
On his return, the mother told him that Billy
was a lot better. As we spoke, said Harry, a
loud hacking cough came from upstairs, and I
commented that he didnt sound better. Oh no,
said the mother, thats not Billy, its Johnny, his
brother. When I oered to take a look at him,
she said, Id rather you didntwe really cant
aord it. Hes just the same as Billy, so Ive given
him some of Billys lehover medicine. It was 5
July 1948 (the hrst day of the NHS). I told her
that from that day it wouldnt cost her anything
and eventually walked away feeling much
lighter in my heart as well as my trouser pocket.
Harry was mutualist to the marrow. He had no
time for the politicians and carpetbaggers who
later sought to make political or hnancial capital
from the disintegration of the NHS. In the 1980s
he forced a High Court review of Kenneth Clarks
introduction of the internal market, and he
was founder and president of the NHS Support
Federation. His passion for justice also made
him the champion of people amicted by the
stigma of chronic disease.
His career spanned the rise of diabetes as a
major public health problem and as a clinical
specialty. It began when he joined George
Pickering, who was at that time measuring the
blood pressure of the population around St
Marys Hospital to establish whether people
with hypertension represented a separate
subgroup or one end of a continuum. Harry was
tasked to measure the blood pressure of people
with diabetes and their families. This early
experience also alerted him to the devastation
caused by the vascular complications of
diabetes, and his compass was set.
From this beginning he, like Kelly West in the
United States, pioneered diabetes epidemiology
before the fact. In 1981 he cofounded the
World Health Organization/International
Diabetes Federations Cambridge seminar on
the epidemiology and public health aspects
of diabetes, and as such he bears much of the
responsibility for the subsequent epidemic of
diabetes epidemiologists. He worked with Robin
Lawrence in the diabetes clinic at Kings College
Hospital, London, followed by a seminal year
with the National Institutes of Health (NIH) in
Bethesda, Maryland. He returned to a lecturers
post at Guys Hospital, London, in 1961,
becoming professor of metabolic medicine in
1971.
When Harry began his career no agreed
dehnition of diabetes existed, and neither was
there an agreed way of measuring glucose
tolerance. Blood glucose, he noted, is (like
blood pressure) continuously distributed in the
population. At what point did a risk factor (the
expression had yet to be invented) turn into a
condition that warranted intervention?
He began by measuring the frequency of
asymptomatic diabetes in the population. In
1962 the population of Bedford was invited to
leave labelled urine samples on their doorsteps,
to be collected by boy scouts over the course of
a weekend. The householder who leh a sample
of sherry was quickly detected and persuaded to
share more of it with the investigators.
Harry progressed from measuring glycaemia
to measuring its consequences in the Whitehall
Harry Keen, professor emeritus of human
metabolism (b 1925; q St Marys Hospital Medical
School 1948; CBE, MD, FRCP), d 5 April 2013.
N
I
T
A

F
O
R
O
U
H
I
Keens experience of an early
diabetes clinicwhere five doctors
or more sat at desks in the same
large room and exchanged muttered
comments with their patientsled
him to conclude that 10 minutes
twice a year was not enough for the
adequate management of diabetes
30 BMJ | 25 MAY 2013 | VOLUME 346
OBITUARIES
Nicholas Priestly
Former general practitioner
Westcliff-on-Sea, Essex (b 1939;
q Leeds University Medical
School 1963), died from chronic
obstructive pulmonary disease on
13 December 2012.
After doing house jobs in Leeds,
Nicholas Priestly (Nick) came to
Westcliff in 1966. He joined with his
two future partners to form a new
group practice in one centralised
surgery. He and his partners steadily
built up a large practice, but it
remained friendly and family oriented.
Nick retired in 1999 after !! years in
the same practice. Until then he had
also been the local medical director
for the deputising service. He worked
in accident and emergency medicine
and administered anaesthesia for
local dentists. Nick enjoyed all sports
and was a doctor for the Southend
Rugby Club. He played golf; enjoyed
walking; and was interested in all
music, bridge, and travelling. He
leaves his wife, Mary; two children;
and three grandchildren.
Tony Bullock
Emanuel Tuckman
Former general practitioner St Pauls
Cray, London (b 1921; q University
College Hospital, London, MD, DCH,
DTM&H), died from cardiac failure
on 31 March 2013.
Emanuel Tuckman (Manny)
volunteered for civilian medical relief
with the Friends Service Unit and United
Nations Relief and Rehabilitation
Administration (UNRRA) in central
China, where he treated hundreds of
children and young adults for visceral
leishmaniasis. On his return he
helped set up the first group general
practice. He was one of the first GP
trainers and was medical officer on the
first Aldermaston march. A Nuffield
travelling fellowship enabled him to
spend six months in the USA, studying
psychosomatic disorders in children.
His doctoral thesis was on backward
readers. He was a keen supporter of the
NHS and never had a private patient.
Predeceased by his wife, he leaves a
daughter and grandson.
Emanuel Tuckman
John T D Attenborough
Former general practitioner
Camberley (b 1920; q Cambridge
1942; MRCS), d 7 March 2013.
Having qualified at the age of ll,
John T D Attenborough joined the
army in 19/!. He served in the
Middle East, Italy, Holland, Germany,
and finally, as staff surgeon, in
Peshawar, India. In 19/', he met
up with his fiance, Frances, after
a separation of two years as she
had been in India and Burma with
the Princess Marys Royal Air Force
Nursing Service (PMRAFNS). They
were married in Peshawar. After
demobilisation in 19/6, John joined
the family practice in Camberley
and continued his military service in
the Territorial Army for 19 years. He
was senior partner of the practice
for 16 years and retired in 19S6.
Predeceased by his wife in luu',
he leaves four children and seven
grandchildren.
John T D Attenborough
Matthew Walter John Boyd
Former consultant physician
Belfast (b 1921; q Queens
University, Belfast, 1943),
d 7 November 2012.
During a long career at the Belfast
City and Musgrave Park hospitals,
Matthew Walter John Boyd promoted
the development of rheumatology
as a specialty in its own right. In due
course his efforts were rewarded
by the appointment of several
consultant rheumatologists in the
various area boards in Northern
Ireland. In his retirement Walter
maintained a continuing interest in
the Irish Society for Rheumatology,
and in September lu1l he received
its lifetime achievement award. He
died after a prolonged period of
deteriorating health, which included
prostate cancer and a series of
small strokes leading to moderately
severe dysphasia. He leaves his
wife, Cathleen (ne Wade), and six
children.
Samuel D Nelson
Derrick Dexter
Former pathologist (b 1923;
q Sheffield 1946; MD, FRCPath,
FRCPC), d 19 March 2013.
Derrick Dexter did his pathology
training at St Georges Hospital
in Hyde Park, London, where he
became a reader in pathology. He
emigrated to Canada with his family
in 196/ and became the director
of laboratories at the Grey Nuns
Hospital, Regina. Apart from a two
year post in New Brunswick, he
spent the rest of his working career
in Regina, as a professor at the
University of Saskatchewan and
director of laboratories at the Plains
Hospital, where his diagnostic skills
were widely sought. He particularly
enjoyed it when his diagnosis on a
rare case was confirmed by the
Mayo Clinic. He and his wife,
Kathleen, returned to the UK in
1991 and retired to Ripon. He
leaves Kathleen, and two sons, and
grandchildren.
David Dexter
Gordon Evison
Consultant radiologist Bath
(b 1933; q Manchester 1957;
DMRD, FRCR), died from a
perioperative stroke on
30 January 2013.
After house appointments in
Manchester, Gordon Evison worked
in internal medicine in hospitals
in Montreal and Vancouver before
returning to train in radiology in
Edinburgh and Bristol. In Bath,
where he worked for !u years,
he was involved in the planning
and development of a new x ray
department at the Royal United
Hospital and an isotope service.
He served as honorary secretary of
the BMAs Bath and district branch
during the 197us and published in
a wide variety of specialty journals,
as well as writing more humorous
pieces. Literature, mature claret,
and Shakespearean theatre were
his lifelong interests. He leaves his
wife, Ruth; four children; and four
grandchildren.
J G Evison
Alexander Iain Munro
Glen
Psychiatrist and scientist (b 1930;
q 1953), died from cancer on
14 February 2013.
Alexander Iain Munro spent his
national service in Hong Kong.
He later trained in psychiatry
in Glasgow and subsequently
developed a lifelong interest in
cell membrane biochemistry and
the importance of the omega !
and omega 6 fatty acids in their
maintenance and function. His
other engrossing passion was
politics, and he made an important
contribution to the Scottish National
Partys national policy development
in the 19Sus and 9us. Latterly Iain
became a carer to his wife, Evelyne,
as her dementia progressed. He
had looked forward to many more
years of reading and writing, but
this was cut short after a diagnosis
of bladder cancer. He leaves
Evelyne; four children; and eight
grandchildren.
Alastair Glen,
Dugald Glen
BMJ | 25 MAY 2013 | VOLUME 346 31
CLINICAL REVIEW
the average person gets clinical malaria four or more
times a year. One observational study found that around
one in hve travellers with a history of fever returning
from sub-Saharan Africa had malaria.
7
Rates are much
lower (maybe 1/100) in unwell travellers returning from
malaria endemic Asia and Latin America, although the
risk of malaria varies widely in these continents.
6
Failure
to take eective prophylaxis is associated with acquiring
malaria in several observational studies. Travellers whose
families emigrated from malaria endemic countries, and
who are visiting friends and relatives, are at particularly
high risk of acquiring malaria, especially those of west
African heritage.
2
People who travelled to visit friends
and relatives account for around 65% of all malaria cases
in the UK, and their reported chemoprophylaxis use is
lower than for other travellers. Vivax malaria, the form
of malaria most commonly imported from Asia and Latin
America, had been declining over the past two decades
in the UK.
8
Observational studies of imported febrile illnesses in
travellers have shown that malaria remains one of the
most common potentially life threatening causes of
fever in travellers.
9
Such studies also show that increas-
ing age, delay to diagnosis,
6
and presenting in an area
where malaria is seldom seen or treated increase the risk
of dying from falciparum malaria once acquired (evidence
from the UK only).
6
Although in Africa most deaths are in
children, in non-endemic high resource settings observa-
tional studies show that older people have the greatest
risk of dying from malaria if acquired, with minimal risk
in children and young adults and a steadily increasing
risk over the age range; case fatality in those over 65 years
is 4.6% in the UK.
5

6
How to recognise malaria in adults
The symptoms and signs of malaria are non-specihc and
overlap with many other common infections. Inuenza is a
common misdiagnosis, and case reports show that malaria
Every year, several thousand people with malaria arrive
in non-endemic countries, and about 1600 arrive in the
United Kingdom alone. Case fatality in the UK, as else-
where, is around 1% overall but varies by age and previ-
ous exposure to malaria.
1-3
This rate is similar to that seen
in endemic countries, but the age prohle of deaths is very
dierent. In Africa mortality is highest in young children,
but in imported cases it is highest in older patients, espe-
cially those over 65 years.
4-6
If malaria is treated early, with
widely available drugs before it becomes severe, death is
avoidable and a full recovery almost guaranteed. Late
presentation carries a higher risk of death. The manage-
ment of severe malaria is a medical emergency. This review
describes how to recognise and diagnose malaria, the cur-
rent treatment of uncomplicated malaria, and the manage-
ment of patients with severe disease. It concentrates on
adults and recent advances relevant to non-endemic high
resource countries such as Europe and North America. Dif-
ferent challenges arise in low resource endemic settings
and are not covered in this review.
What is malaria?
Malaria is a tropical parasitic disease of red blood cells
transmitted by anopheles mosquitoes. Five species aect
humans. The most serious, and in Europe the most com-
mon, imported form is Plasmodium falciparum malaria,
which causes most deaths from imported malaria. The
other common imported form is P vivax malaria. It is
generally less severe but can recur several months aher
treatment owing to relapse from hypnozoites (sleeping
parasites) in the liver. P ovale malaria is similar to P vivax
malaria, as is P malariae malaria although it does not
relapse. P knowlesi is a monkey malaria parasite that has
recently been recognised to infect humans and is rare in
imported cases.
Who is at risk of acquiring and dying from malaria?
Several countries that are frequented by tourists and
people visiting friends and relatives have regions where
1
Hospital for Tropical Diseases,
London WC1E 6JB, UK
l
PHE Malaria Reference Laboratory,
London School of Hygiene and
Tropical Medicine, London, UK
!
Liverpool School of Tropical
Medicine, Liverpool, UK
Correspondence to: C J M Whitty
christopher.whitty@lshtm.ac.uk
doi: 1u.11!6/bmj.fl9uu
Investigation and treatment of imported
malaria in non-endemic countries
Christopher J M Whitty,
1 l
Peter L Chiodini,
1 l
David G Lalloo
!
SUMMARY POINTS
Malaria is a common cause of fever in people returning
from the tropics
Falciparum malaria is potentially fatal unless treated early,
and patients over 65 years are at particular risk
In most cases, vivax malaria can be treated with
chloroquine in the outpatient setting
Resistance to antimalarial drugs used to treat falciparum
malaria is widespread
Artesunate is the drug of choice for severe malaria, but if
this is not available do not delay treatment with quinine
Follow the linkfrom the
online version of this article
to obtain certied continuing
medical education credits
SOURCES AND SELECTION CRITERIA
We searched the Cochrane Library and PubMed for recent
relevant trials and observational studies on imported
malaria. The search was supplemented where relevant by
data from the UK Malaria Reference Laboratory and expert
opinion from the PHE Advisory Committee on Malaria
Prevention in UK Travellers (ACMP) and World Health
Organization expert guidelines. Good epidemiological
data are available on imported malaria in Europe and North
America and on diagnostic tests. There are high quality trials
of malaria treatment in endemic countries but few large
well conducted trials of treatment in travellers, so data on
treatment have to be extrapolated from endemic settings.
bmj.com
Previous articles in
this series
Safeguarding adults
at risk of harm
(BMJ lu1!;!/6:fl716)
Acne vulgaris
(BMJ lu1!;!/6:fl6!/)
Adolescent idiopathic
scoliosis
(BMJ lu1!;!/6:fl'u3)
Diagnosis and
management of
hidradenitis suppurativa
(BMJ lu1!;!/6:fl1l1)
Pulmonary
hypertension: diagnosis
and management
(BMJ lu1!;!/6:flul3)
32 BMJ | 25 MAY 2013 | VOLUME 346
CLINICAL REVIEW
has also been misdiagnosed as gastroenteritis, hepatitis,
and lower respiratory tract infection.
10
The key to diagnos-
ing malaria is to take a travel history for all patients and
request a malaria test in anyone who feels unwell and has
recently been to an endemic country. A history of fever will
be present in most patients with malaria, but around half
will have no fever at the time of presentation; absence of
fever cannot exclude disease.
7
Headache, malaise, and
rigors are common symptoms but not invariable. Certain
symptoms and signs of severe or potentially complicated
malariasuch as jaundice, seizures, or rapid breathing
can result in the misdiagnosis of malariafor example,
as hepatitis (table 1). It is therefore essential to exclude
malaria with a thick and thin blood hlm rather than rely
on clinical diagnosis based on symptoms; attempts to pro-
duce algorithms that reliably predict malaria on the basis
of symptoms have been unsuccessful.
11
In patients return-
ing from malaria endemic countries, any symptom or sign
compatible with an infection should raise the possibility
of malaria. Although most cases of falciparum malaria
present within three months of return, case reports and
Public Health England data show that some cases in trav-
ellers present up to a year aher return, and many cases of
non-falciparum malaria occur over a year later.
12
How is malaria diagnosed?
The gold standard for malaria diagnosis in clinical practice
remains microscopic examination of a blood hlm. Good
rapid diagnostic tests for falciparum malaria with sen-
sitivity and specihcity above 97%, and fairly good ones
for non-falciparum malaria,
13
are now widely available to
hospitals in Europe. Most, however, are not as sensitive
as an experienced microscopist and give no information
about parasite density, which is important for prognosis
and treatment decisions. Therefore these rapid tests should
be seen as an adjunct to, rather than an alternative to, con-
ventional microscopic diagnosis, and their sole use is not
recommended. Newer polymerase chain reaction based
diagnostic methods are currently used only in research
settings, although they will probably become clinically
useful in non-endemic settings within the next decade,
especially to detect mixed infections and potentially drug
resistant cases.
14
What are the indicators of severe or complicated disease
in adults?
Once malaria is diagnosed, the priority is to determine
the severity of disease because this determines treatment.
Table 1 shows modihed World Health Organization indi-
cators of severe disease; patients with any one of these
will need parenteral treatment with antimalarial drugs.
HIV positive patients have an increased risk of severe
malaria and of drug interactions.
15
Observational stud-
ies have shown that three severe syndromes predominate
in adults: cerebral malariain practice, malaria with
reduced level of consciousness, neurological signs, or
seizures; renal failure; and acute lung injury (or acute
respiratory distress syndrome).
16
Disseminated intra-
vascular coagulation and shock are less common and
seldom occur in isolation. In addition, jaundice (due to
recent major haemolysis) or a parasite count in the blood
of greater than 2%, even if the patient does not appear
clinically unwell, is a sign that severe malaria is more
likely to develop.
What are the indicators of severe or complicated disease
in children?
Multicentre observational studies, mainly from Africa,
have shown that anaemia, raised respiratory rate due to
acidosis (rather than lung injury), and cerebral malaria
are the most common signs of severity in young children,
although this pattern varies by age, with cerebral malaria
becoming more common in older children.
17
What about malaria in pregnancy?
Treat pregnant women as having potentially complicated
malaria, although not all will need parenteral treatment.
The risks associated with severe disease are greater than
in non-pregnant women, and there are serious risks to
the pregnancy, including miscarriage, in otherwise non-
severe disease.
18
The management of malaria in preg-
nancy has particular challenges,
19
because of the risks
the disease poses to the pregnancy and because the theo-
retical risks of teratogenicity of antimalarial drugs need to
be balanced with the substantial risks of undertreating;
specialist advice is strongly recommended. Data from
animals suggest that artemisinin drugs are teratogenic
in early pregnancy, but human data are reassuring, with
no current evidence of increased risk.
20
Hypoglycaemia
is a problem in pregnant women with malaria,
21
and any
pregnant woman with a reduced level of consciousness
should be given glucose while awaiting blood glucose
measurements.
How is uncomplicated non-falciparum malaria managed?
Several case reports indicate that vivax malaria, and
possibly ovale and malariae malaria, cause severe dis-
ease more ohen than was once thought,
22
and their label
as benign malaria is misleading. However, patients
with non-falciparum malaria need to be admitted only
if they have indicators of severe disease (table 1) or can-
not take oral drugs. Provided the laboratory making the
Table | WHO criteria for severe malaria, modified for non-
endemic high resource settings (any one indicates potentially
severe disease)*
Symptoms, signs, and laboratory indicators of
severe disease Adults Children
Reduced level of consciousness +++ +++
Seizures + ++
Respiratory distress ++ ++
Severe anaemia (haemoglobin < g/L)* + +++
Hypoglycaemia + ++
Renal failure +++ +
Hyperparasitaemia (>% red cells parasitised) ++ ++
Disseminated intravascular coagulation + +
Shock + +
Warning signs of potentially complicated malaria (requires parenteral
treatment):
Jaundice + +
Parasite count >% +++ +
*+=rare; ++=uncommon; +++=common (definition depends on age, state of
immunity, and other factors).
Common in pregnant women.
BMJ | 25 MAY 2013 | VOLUME 346 33
CLINICAL REVIEW
diagnosis is experienced and unlikely to have mistaken
falciparum or rare mixed species infection (<1%), chlo-
roquine remains the drug of choice to treat vivax, ovale,
and malariae malaria. There is limited but growing
chloroquine resistance in vivax in eastern Asia and the
Pacihc.
23
Evidence from clinical trials and a systematic
review shows that most patients with vivax malaria, even
those from South Asia or East Asia, respond to chloro-
quine within three days (see table 2 for dosing).
24
Good
evidence from clinical trials shows that artemisinin com-
bination treatment (ACT) or quinine both work against
vivax malaria at the same doses as those used for falci-
parum malaria (see below) if there is uncertainty about
species or in patients with mixed infections.
25
The biggest dimculty in managing patients with vivax
or ovale malaria is that they can relapse several times
aher an initial episode, ohen more than a year aher initial
infection. Chloroquine and ACT treat the initial infection
but do not kill the hypnozoites in the liver, which cause
recurrent malaria. Primaquine is the only drug currently
licensed to kill these hypnozoites and prevent or reduce
the risk of relapse.
26
Apart from the need for a two week
course, primaquine has two disadvantages. Firstly, resist-
ance is gradually spreading, particularly in Oceania and
eastern Asia, so that higher doses are now needed.
27
In
the United States and UK, the current recommended dose
for vivax malaria in adults is 30 mg a day for 14 days (15
mg a day for ovale). In addition, case reports and observa-
tional studies have found that treatment with primaquine
in patients with phenotypic glucose-6-phosphate dehy-
drogenase (G6PD) dehciency can lead to haemolysis,
which can be life threatening. G6DP dehciency therefore
needs to be excluded before the drug is taken.
28
G6PD
dehciency can aect more than 10% of the population
in malaria endemic countries, although its prevalence
is lower in people with malaria.
29

30
Patients should not
take primaquine until their G6PD status is known, but
treatment with chloroquine must not be delayed. Unfor-
tunately, the only new anti-hypnozoite drug likely to be
licensed in the next few years, tafenoquine, has similar
risks with G6PD dehciency.
How should uncomplicated falciparum malaria be managed?
The severity of uncomplicated falciparum malaria is dif-
hcult to assess at presentation owing to the complex life
cycle of the parasites, so all patients presenting with
falciparum malaria in non-endemic countries should be
admitted. Patients can seem well initially and then deterio-
rate despite treatment, and in around 20% of patients the
parasite count rises in the hrst 24 hours despite adequate
treatment.
7
Do not assume that patients who once lived
in endemic countries but now are settled in non-endemic
countries retain immunity. Clinical studies show that
although being born in endemic Africa reduces the risk of
dying from malaria, admission to intensive care and death
can still occur.
6

31
Provided oral treatments can be tolerated, parenteral
treatment is not needed if there are no clinical or laboratory
signs of severity (table 1). Because of a lack of evidence,
guidelines vary about the threshold parasite count above
which to give parenteral treatment in patients without
symptoms or signs of severity: UK guidelines (designed for
non-endemic countries) suggest parasite counts over 2%,
whereas WHO suggests over 5%. Chloroquine resistance is
near universal, so this drug should not be used. A variety
of drug combinations are eective against uncomplicated
falciparum malaria (table 2). Falciparum malaria should
always be treated with a combination of at least two drugs.
Commonly used options include ACTs, quinine combined
with another drug, and atovaquone-proguanil. Most non-
endemic countries have evidence based expert recom-
mendations that are periodically updated on the basis of
trends in epidemiology of imported malaria, emerging drug
resistance, and local availability of drugs.
32

33
There is good evidence from large trials and systematic
reviews that ACT or quinine based combinations will be
eective in malaria from Africa, most of Asia, and Latin
America; ACT drugs are generally better tolerated and reduce
parasite counts more rapidly.
34
Two ACTs, artemether-lume-
fantrine and dihydroartemisinin-piperaquine, are currently
licensed in Europe and may become available in the US.
Is malaria resistance increasing?
P falciparum has developed resistance to several older
antimalarials, and resistance continues to evolve. Sev-
eral clinical studies show clear evidence of early emerg-
ing artemisinin resistance in South East Asia (Cambodia,
Thailand, and parts of Burma), although ACTs still work
clinically.
35
Clinical trials have also noted partial quinine
resistance in the same area. Several recent trials have
found no evidence that quinine or artesunate resistance
is a serious clinical problem outside these areas. Resist-
ance to some older companion drugs (such as sulfadoxine-
pyrimethamine) is widespread, but well conducted trials
(in endemic countries) and observational data (in imported
cases) show that the combinations in table 2 are eective
for most (>95%) cases of imported malaria. Atovaquone-
proguanil has been associated with occasional treatment
failure in some case reports, so many centres do not use it
Table | Common licensed drugs, drug combinations, and drug doses for malaria
Drug Dose in adults Notes
Uncomplicated vivax, ovale, and malariae malaria
Chloroquine 6lu mg base, then !1u mg base at 6, l/,
and /S hours
Primaquine (after chloroquine) !u mg per day for 1/ days Test for G6PD deficiency before
patients take it
Uncomplicated falciparum malaria
Artemether-lumefantrine / tablets (adult) initially, followed by '
further doses of / tablets at S, l/, !6, /S,
and 6u hours
Quinine combinations 6uu mg every S hours for 7 days or until
parasites have cleared
Follow with clindamycin for 7
days, doxycycline for 7 days, or
sulfadoxine-pyrimethamine once;
ringing in ears a common side effect
Atovaquone-proguanil / tablets once a day for ! days Avoid if patients were on
atovaquone-proguanil prophylaxis
Severe or potentially complicated malaria
Quinine lu mg/kg quinine intravenously, followed
by 1u mg/kg S-1l hourly until parasites
have cleared
By slow infusion over / hours;
hypoglycaemia and arrhythmias are
side effects
Artesunate l./ mg/kg intravenously, followed by l./
mg/kg at 1l hours, then l./ mg/kg daily
until parasites have cleared
As a bolus over ' minutes
G6PD=glucose-6-phosphate dehydrogenase.
34 BMJ | 25 MAY 2013 | VOLUME 346
CLINICAL REVIEW
quinine, artesunate kills parasites of all stages and rapidly
reduces parasite counts. Exchange transfusion is now rarely
indicated and should be used only aher discussion with a
specialist centre.
The optimum management of uids in adults and chil-
dren has, in the absence of good data, been highly controver-
sial. A recent trial in African children showed convincingly
that aggressive fluid management was associated with
increased mortality.
41
Good data on the optimal manage-
ment of uids in adults or children in high resource settings
are limited. In adults, particularly those with renal failure,
the risks of undertreating acidosis and any pre-renal com-
ponent if uids are withheld must be balanced against pro-
voking pulmonary oedema if excess uids are used. Most
doctors used to treating severe malaria are wary about the
overuse of uids once pre-renal failure has been ruled out
by short, small uid challenges; unlike in bacterial sepsis,
shock is rare in adults with malaria.
Clinical observational studies have shown a clear asso-
ciation between severe malaria and Gram negative sepsis
in children, and reductions in the incidence of malaria in
populations have been associated with reductions in Gram
negative bacteria.
42

43
Broad spectrum antibiotics should
therefore be considered in all children with severe malaria.
Data from adults are limited, but the association seems to
be less pronounced, so routine antibiotics are not needed.
44

Case reports suggest that broad spectrum antibiotics should
be given to the rare adult patients who have shock with
malaria.
Acute respiratory distress syndrome is the most feared late
complication of malaria in adults. In severe malaria, this
can occur several days aher treatment has started, including
when parasites have cleared from the blood. As with other
causes of the syndrome, it is important to maintain oxygena-
tion with respiratory support in intensive care and treat the
underlying infection. No adjunctive treatments have been
shown to improve outcomes in patients with this compli-
cation. Respiratory distress is also a poor prognostic sign
in African children,
11
but less so in Asia (perhaps because
of greater availability of blood transfusions)
45
; it is almost
always caused by acidosis rather than lung injury.
11
What is the prognosis after malaria?
The outlook for adults who survive an initial episode of
malaria is good, even for those with severe malaria. Serious
neurological sequelae occur in less than 5% of adults with
severe malaria, although few series have been large enough
to provide conhdent estimates. Many of the estimates in the
literature are based on expert opinion rather than data.
46

47

Renal failure and lung injury almost invariably resolve if
patients survive. Neurological dehcits in children who have
had cerebral malaria are now recognised to be more com-
mon, up to 30% in some series,
48
but these are ohen more
subtle than for other serious neurological infections, such
as meningitis.
49
PLC is supported by the UCL Hospitals Comprehensive Biomedical
Research Centre Infection Theme.
Contributions: All authors contributed to the writing of this paper. CW is
guarantor.
Provenance and peer review: Commissioned; externally peer reviewed.
as hrst line treatment
36
; it should not be used in people who
have been taking it prophylactically. Atovaquone-proguanil
may be appropriate second line treatment for the rare cases
of treatment failure with eective hrst line combinations,
particularly those from South East Asia. Falciparum malaria
does not relapse and patients can be reassured that they
are unlikely to have recurrent malaria if they complete a full
course of treatment. Any further attacks (generally within six
weeks) usually represent treatment failure. Warn travellers
about the growing problem of poor quality and counterfeit
antimalarial drugs in many endemic countries.
37
How should severe and potentially complicated malaria be
managed?
Severe malaria is a medical emergency, and the main prior-
ity is for patients to receive adequate doses of eective drugs
as soon as possible. Quinine (quinidine in the US, which is
broadly equivalent) or artesunate are the two drugs used for
parenteral treatment of severe malaria. Clear evidence from
large randomised trials now shows that although quinine
remains eective, artesunate is associated with a survival
advantage (relative risk reduction of 22-34%) in adults and
children in Asia and Africa.
38

39
Previously, the poor quality
and limited availability of parenteral artesunate has been a
problem in many non-endemic countries. However, although
not licensed in Europe or the US, parenteral artesunate of
reliable quality can now be sourced (including through
the Centers for Disease Control and Prevention in the US).
Patients should always be started on parenteral quinine
while trying to obtain artesunate if it is not initially available
to avoid delay in eective treatment. In adults, artesunate
confers the greatest survival beneht in those with very high
parasite counts (>10%).
22
Parenteral quinine and quinidine
are associated with hypoglycaemia and arrhythmia so should
be used with caution in patients with cardiac problems.
Other than antimalarials, good management of patients
with severe malaria largely depends on supportive care,
including dialysis or haemohltration in those with renal fail-
ure and respiratory support for those with acute lung injury.
Various adjunctive treatments have been tried in cerebral
malariaincluding steroids, mannitol, and anti-tumour
necrosis factornone of which has shown a survival advan-
tage.
For many years, there has been controversy over the use of
exchange transfusion or automated red blood cell exchange
to reduce parasite counts in patients with hyperparasitaemia
(generally >10%).
40
The advent of treatment with artesu-
nate, when available, renders this debate irrelevant. Unlike
ADDITIONAL EDUCATIONAL RESOURCES
Resources for healthcare professionals
WHO World Malaria Report 2u12. www.who.int/malaria/publications/world_malaria_
report_2u12/en/. Describes the current state of malaria in the world
WHO management of severe malaria: a practical handbook. 2u13. www.who.int/malaria/
publications/atoz/97892/15/8526/en/index.html. Explains the management of severe
malaria
Resources for patients
National Travel Health Network and Centre. www.nathnac.org/travel/factsheets/malaria_
chemoprophylaxis.htm. Information leaflet for travellers on chemoprophylaxis
Advisory Committee on Prevention of Malaria in UK. www.hpa.org.uk/infections/topics_az/
malaria/guidelines.htm. Travellers handbook
BMJ | 25 MAY 2013 | VOLUME 346 35
PRACTICE
Colour (of skin,
lips, or tongue)
Activity
Respiratory
Circulation and
hydration
Other
This tramc light table should be used in conjunction with the recommendations in this guideline on investigations and
initial management in children with fever
AmberIntermediate risk RedHigh risk GreenLow risk
Normal colour
Responds normally to social
cues
Content or smiles
Stays awake or awakens
quickly
Strong normal cry or not
crying
Normal skin and eyes
Moist mucous membranes
None of the amber or red
symptoms or signs
Pallor reported by parent or
carer
Not responding normally to social
cues
No smile
Wakes only with prolonged
stimulation
Decreased activity
Nasal flaring
Tachypnoea:
Respiratory rate (RR) >5u
breaths/min at ages 612 months
RR >/u breaths/min at ages
>12 months
Oxygen saturation 95% in air
Crackles in the chest
Tachycardia:
>16u beats/min at age <12 months
>15u beats/min at age 12-2/ months
>1/u beats/min at age 2-5 years
Capillary rell time 3 seconds
Dry mucous membranes
Poor feeding in infants
Reduced urine output
Temperature 39C at ages
36 months
Fever for 5 days
Rigors (see denitions box)
Swelling of a limb or joint
Non-weight bearing limb or not
using an extremity
Pale, mottled, ashen, or blue
No response to social cues
Appears ill to a healthcare
professional (see box of
denitions of terms)
Does not wake, or if roused
does not stay awake
Weak, high-pitched, or
continuous cry
Grunting
Tachypnoea:
RR > 6u breaths/min
Moderate or severe chest
indrawing
Reduced skin turgor
Temperature 38C at ages
<3 months
Non-blanching rash
Bulging fontanelle
Neck stinness
Status epilepticus
Focal neurological signs
Focal seizures
National Collaborating Centre for

Womens and Childrens Health,
London WT QA, UK
Peterborough and Stamford

Hospitals Foundation Trust,
Peterborough PE GZ, UK
Correspondence to: J Chard
jchard@ncc-wch.org.uk
doi: ./bmj.f
This is one of a series of BMJ
summaries of new and updated
guidelines based on the best
available evidence; they highlight
important recommendations for
clinical practice, especially where
uncertainty or controversy exists.
Further information about the
guidance, a list of members of the
guideline development group, and
the supporting evidence statements
are in the full version on bmj.com.
Among children presenting with fever, especially in a
primary care setting, serious illness is uncommon. The
prevalence of serious illness has been reported at .%
in primary care
and .% in secondary care.
For this
reason it is important that guidance is available to help
healthcare professionals distinguish the many who have
minor transient conditions from the occasional child with
a serious or even life threatening infection. This article
summarises the key recommendations from the
update of the National Institute for Health and Care Excel-
lence (NICE) guidelines on feverish illness in children.
Recommendations
NICE recommendations are based on systematic reviews
of the best available evidence and explicit considera-
GUIDELINES
Assessment and initial management of feverish illness in children
younger than 5 years: summary of updated NICE guidance
Ella Fields,
1
Jiri Chard,
1
M Stephen Murphy,
1
Martin Richardson,
2
on behalf of the Guideline
Development Group and technical team
tion of cost-effectiveness. When minimal evidence is
available, recommendations are based on the Guideline
Development Groups (GDG) experience and opinion of
what constitutes good practice. Evidence levels for the
recommendations are in the full version of this article on
bmj.com.
Clinical assessment of feverish children

This should consist of three stages:

Identify life threatening features (airway, breathing,
circulation, disability). If any are present, refer
immediately for emergency medical care.

Use the trac light table (g ) to assess the risk
of serious illness as being low (green), intermediate
(amber), or high (red). (Updated recommendation
Fig | Traffic light table
for assessing risk of serious
illness in feverish children
aged < years old. Children
with fever and any features
from the red column should
be considered as being at
high risk. Children with fever
and any features from the
amber column and none in
the red column should be
considered at intermediate
risk. Children with features in
the green column and none in
the amber or red columns are
at low risk. (This table should
be used only in conjunction
with this guideline on
investigations and initial
management)
36 BMJ | 25 MAY 2013 | VOLUME 346
PRACTICE

Rigors was added to the amber column

Age 3-6 months, temperature 39C was moved
from the red (high risk) column to the amber column

A new lump >2 cm was removed

Bile-stained vomiting was removed.
Antipyretic agents

Consider using either paracetamol or ibuprofen in
children with fever who seem distressed. (Updated
recommendation.)

Do not use antipyretic agents with the sole aim of
reducing body temperature in children with fever.
(Updated recommendation.)

Antipyretic agents do not prevent febrile convulsions
and should not be used just for this purpose.

When using paracetamol or ibuprofen in children with
fever:

Continue only as long as the child appears distressed

Consider changing to the other agent if the childs
distress is not alleviated

Do not give both agents simultaneously

Consider alternating these agents only if the distress
persists or recurs before the next dose is due.
(Updated recommendation.)
with substantive changes, including addition of
tachycardia as a risk factor for serious illness.)

Attempt to identify a focus of infection or features of
specihc serious conditions (see table).

Measure and record temperature, heart rate,
respiratory rate and capillary rehll time as part of the
routine assessment of a child with fever.

In children aged 4 weeks to 5 years, measure body
temperature by one of the following methods:

Electronic thermometer in the axilla

Chemical dot thermometer in the axilla

Infra red tympanic thermometer

Reported parental perception of a fever should be
considered valid and taken seriously by healthcare
professionals.
Management
Management in primary care (hg 2) and specialist care (hg
3) is determined by the assessment of risk of serious illness.
The following substantive changes have been made for
the 2013 update:

Tachycardia was added to the amber (intermediate
risk) column (age related values from Advanced
Paediatric Life Support
6
are appropriate)
Clinical features of specific serious diseases in conjunction with fever
Diagnosis to be considered Symptoms and signs in conjunction with fever
Meningococcal disease* Non-blanching rash, particularly with 1 of the following:
Ill looking child Lesions >l mm in diameter (purpura) Capillary refill time ! seconds Neck stiffness
Bacterial meningitis* Neck stiffness Bulging fontanelle Decreased level of consciousness Convulsive status epilepticus
Herpes simplex encephalitis Focal neurological signs Focal seizures Decreased level of consciousness
Pneumonia Tachypnoea (respiratory rate >6u breaths/min at ages u' months, >'u breaths/min at 61l months, or >/u breaths/min
at >1l months) Cyanosis Nasal flaring Chest indrawing Crackles in the chest Oxygen saturation 9'%
Urinary tract infection Lethargy Irritability Vomiting Poor feeding Abdominal pain or tenderness
Urinary frequency or dysuria
Septic arthritis Not using an extremity Non-weight bearing Swelling of a limb or joint
Kawasaki disease Fever for >' days and / of the following:
Bilateral conjunctival injection Change in mucous membranes (such as injected pharynx, dry cracked lips, or strawberry
tongue) Change in extremities (such as oedema, erythema, or desquamation) Polymorphous rash Cervical
lymphadenopathy
*See NICE guideline on bacterial meningitis and meningococcal septicaemia for more detail.
/
See NICE guideline on urinary tract infection in children for more detail.
'
Children with only green features
can be cared for at home with
appropriate advice for parents and
carers, including advice on when to
seek further attention from
healthcare services
If any amber features are present
and no diagnosis has been reached,
provide parents or carers with a
safety net (see box of denitions
of terms) or refer to specialist
paediatric care for further
assessment. The safety net should
be of the following:
Providing verbal and/or written
information on warning symptoms
and how further healthcare can be
accessed
Arranging further follow-up at a
specied time and place
Liaising with other healthcare
professionals, including out of hours
care providers, to ensure direct
access for the child if further
assessment is required
For children whose clinical features
suggest an immediately life
threatening illness, refer
immediately for emergency medical
care by the most appropriate means
of transport
From remote assessmentChildren
with any red features but not
considered to have an immediately
life threatening illness should be
urgently assessed face to face by a
healthcare professional within
hours
From non-specialist careFor
children with any red features but
not considered to have an
immediately life threatening illness,
refer urgently to the care of a
paediatric specialist
Fig | Management of
feverish illness in children
aged < years old by the non-
specialist
bmj.com
this series
Long term follow-up
of survivors of childhood
cancer: summary of
updated SIGN guidance
(BMJ lu1!;!/6:f119u)
Recognition,
intervention, and
management of
antisocial behaviour
and conduct disorders
in children and young
people: summary of
NICE-SCIE guidance
(BMJ lu1l;!/6:f1l93)
Fertility (update):
summary of NICE
guidance
(BMJ lu1!;!/6:f6'u)
Recognition
and management
of psychosis and
schizophrenia in children
and young people:
summary of NICE
guidance
(BMJ lu1!;!/6:f1'u)
Ectopic pregnancy and
miscarriage: summary of
NICE guidance
(BMJ lu1l;!/':e31!6)
BMJ | 25 MAY 2013 | VOLUME 346 37
PRACTICE
the tramc light table in isolation.
10
For example, the Yale
Observation Scale is moderately useful at identifying seri-
ous illness but could not be used as a rule out criterion.
It is important that the tramc light table in this guideline
is used in combination with other recommendations,
both in this guideline (such as the need for urine analysis)
and in other relevant guidelines (such as the NICE Clini-
cal Guidelines on urinary tract infection in children
5
and
bacterial meningitis and meningococcal septicaemia
4
) for
complete management.
Contributors: All authors contributed to the initial drahing of this article
and revising it critically. They have all approved this version. MSM is the
guarantor.
Competing interests: We have read and understood the BMJ Group
policy on declaration of interests and declare the following interests: EF,
JC, and MSM have support from the National Institute for Health and Care
Excellence for the submitted work.
Provenance and peer review: Commissioned; not externally peer
reviewed.
1 Van den Bruel A, Aertgeerts B, Bruyninckx R, Aerts M, Buntinx F. Signs
and symptoms for diagnosis of serious infections in children:
a prospective study in primary care. Br J Gen Pract luu7;'7:
'!S-/6.
l Craig JC, Williams GJ, Jones M, Codarini M, Macaskill P, Hayen A, et
al. The accuracy of clinical symptoms and signs for the diagnosis
of serious bacterial infection in young febrile children: prospective
cohort study of 1' 7S1 febrile illnesses. BMJ lu1u;!/u:c1'9/.
! National Institute for Health and Care Excellence. Feverish illness in
children: assessment and initial management in children younger
than five years. (Clinical guideline CG16u.) lu1!. http://guidance.
nice.org.uk/CG16u).
/ National Institute for Health and Clinical Excellence. Bacterial
meningitis and meningococcal septicaemia: management of bacterial
meningitis and meningococcal septicaemia in children and young
people younger than 16 years in primary and secondary care. (Clinical
guideline 1ul.) lu1u. http://guidance.nice.org.uk/CG1ul.
' National Institute for Health and Clinical Excellence. Urinary tract
infection: diagnosis, treatment and long-term management of urinary
tract infection in children. (Clinical guideline '/.) luu7. http://
guidance.nice.org.uk/CG'/.
6 Advanced Life Support Group. Advanced paediatric life support: the
practical approach. /th ed. BMJ Books/Blackwells, luu'.
7 National Institute for Health and Clinical Excellence. Feverish illness
in children: assessment and initial management in children younger
than five years. (Clinical guideline /7.) luu7. http://guidance.nice.
org.uk/CG/7.
S Harnden A. Recognising serious illness in feverish young children in
primary care. BMJ luu7;!!':/u9.
9 De S, Williams GJ, Hayen A, Macaskill P, McCaskill M, Isaacs D, et al.
Accuracy of the traffic light clinical decision rule for serious bacterial
infections in young children with fever: a retrospective cohort study.
BMJ lu1!;!/6:fS66.
1u Thompson M, Van den Bruel A, Verbakel J, Lakhanpaul M, Haj-Hassan
T, Stevens R, et al. Systematic review and validation of prediction
rules for identifying children with serious infections in emergency
departments and urgent-access primary care. Health Technol Assess
lu1l;16(1'):1-1uu.
Overcoming barriers
Parents and carers are naturally worried by feverish ill-
ness in their children, and ohen the initial thought is
to reduce the temperature using products containing
ibuprofen or paracetamol. This guideline makes it clear
that ibuprofen and paracetamol should be used only to
reduce distress in a child, rather than to reduce tempera-
ture alone. It is important for health professionals to get
this message across to parent and carers, and a leaet
has been produced to support this (http://publications.
nice.org.uk/ifp160).
The original tramc light system for assessing risk of
serious illness was generally well accepted in the National
Health Service,
7
but, there has been some concern that it
was not validated.
8
However, a recent article goes some
way towards validating it,
9
and its data were incorporated
in this 2013 update.
2
A recent Health Technology Assessment report has
highlighted potential problems of using tools such as
DEFINITIONS USED IN GUIDELINE
FeverDefined as an elevation of body temperature
above the normal daily variation for the purposes of this
guideline.
Appears ill to a healthcare professionalAn overall
impression the assessing healthcare professional
makes when a child presents. This impression is formed
not only from objective measurements but also from
subjective feelings about how the child looks or reacts.
If a healthcare professionals subjective impression is
that the child is ill looking, then the child is probably at
high risk of serious illness.
Healthcare professionals
should be confident to follow their impressions of a childs
wellbeing.
RigorsAn episode of shaking or shivering, which can
occur when the child has high temperature. Rigors can
be confused with febrile convulsions. However, unlike
in a seizure, the child is conscious and alert during the
episode.
Safety nettingProviding support for a patient when the
clinician is uncertain as to the presence of a self limiting
illness and is concerned that the patients condition may
deteriorate. Safety netting may take different forms, such
as dialogue with the parent or carer about symptoms
and signs to watch for, advice about when to seek further
medical attention, review after a set period, and liaising
with other healthcare services.
Test urine for urinary tract infection
Check for specic symptoms and
signs of pneumonia
Do not routinely perform blood tests
or chest x rays in children with fever
who have no amber or red features
of serious illness
The following tests may be
performed:
Urine testing for urinary tract
infection
Full blood count, C reactive protein,
and blood cultures
Consider lumbar puncture for
children < year old
Chest x ray in a child with a fever
>C and leucocyte count
>
/L
Undertake the following tests:
Full blood count
Blood culture
C reactive protein
Urine testing for urinary tract
infection
Consider the following
investigations:
Lumbar puncture in children of all
ages (if not contraindicated)
Chest x ray irrespective of body
temperature and leucocyte count
Serum electrolytes
Blood gases
Fig | Management of
feverish illness in children
aged < years old by the
paediatric specialist
38 BMJ | 25 MAY 2013 | VOLUME 346
PRACTICE
A 35 year old woman presents to her general practitioner
with a 10 day history of worsening nasal congestion,
purulent nasal discharge, and frontal headaches.
What you should cover
Ask about
Rhinosinusitis (including nasal polyps) is an inamma-
tory condition of the nose and paranasal sinuses. Diagno-
sis requires at least two symptoms, one of which must be
nasal discharge or obstruction, with the others compris-
ing facial pain or smell disturbance.
Acute rhinosinusitis is dehned as symptoms lasting
less than 12 weeks with complete resolution and can be
subdivided into

Acute viral rhinosinusitis (common cold)Dehned
by duration of symptoms of less than 10 days.

Acute non-viral rhinosinusitisDehned by an
increase in symptoms aher 5 days or persistent
symptoms aher 10 days.

Acute bacterial rhinosinusitisSuggested by the
presence of at least three symptoms or signs of

Discoloured discharge (with unilateral
predominance) and purulent secretions

Severe local pain (with unilateral predominance)

Fever (>38C)

Elevated erythrocyte sedimentation rate or C
reactive protein

Double sickening (that is, a deterioration aher
an initial milder illness).
Chronic rhinosinusitis (with or without polyps) is
dehned as more than 12 weeks of symptoms without
complete resolution.
The following points are important within the history
in acute rhinosinusitis:

Nasal blockage, obstruction, or congestionIs the
blockage unilateral or bilateral? Acute rhinosinusitis
is usually associated with bilateral symptoms. With
unilateral symptoms, bear in mind the possibility
(albeit rare) of an underlying malignancy.

Nasal discharge (anterior or posterior nasal drip)
Attempts should be made to assess and record the
character, amount, and pattern of nasal discharge
over time.

Facial pain or pressureFacial pain without nasal
obstruction or discharge is highly unlikely to be due
to sinusitis. Purely unilateral facial pain is unlikely to
be sinogenic, most commonly it is dental in origin.

Change, reduction, or loss of sense of smell.

Resolution of symptomsFrequent short lasting
episodes with complete resolution are likely to be
associated with acute viral rhinosinusitis, whereas
infrequent long lasting episodes with no resolution
are suggestive of chronic rhinosinusitis.

Respiratory symptomsThese may include
pharyngeal, laryngeal, or tracheal irritation causing
sore throat, change in voice, and cough.

Systemic symptomsMalaise, headache, and fever
may also occur.
Examination

If the patient seems to be systemically unwell then
heart rate, blood pressure, and temperature should
be assessed. A fever of >38C is more likely to be
associated with bacterial infection.

Percussion over the maxillary, ethmoid, and frontal
sinuses or leaning forward may exacerbate facial
pain or pressure. However, the sensitivity and
specihcity of these signs in the identihcation of acute
bacterial rhinosinusitis have not been established,
and they are not diagnostic in isolation.

Perform anterior rhinoscopy (with an otoscope
or Thudichums nasal speculum with headlight,
depending on availability) and look for

Mucopurulent discharge or nasal polyps (polyps
can sometimes be confused with an engorged
inferior turbinate, but they are insensate, in
contrast to the inferior turbinate)

Other nasal pathology (such as a neoplasm,
particularly in the presence of unilateral polyp or
mass and associated bloody nasal discharge)

In case of diagnostic doubt (that is, symptoms
suggestive of neoplasm as mentioned above),
patients should be referred for nasal endoscopy
(rigid or hbreoptic), which is currently the
optimum method for nasal examination.

Imaging:

A plain radiograph of the sinuses is not
recommended for the diagnosis and management
of acute rhinosinusitis

Computed tomography is the primary investigation
carried out by otolaryngologists if complications
are suspected or when planning endoscopic sinus
surgery.
What you should do
Management (see figure)
Acute rhinosinusitis is common, with an annual p revalence
DIFFERENTIAL DIAGNOSIS OF ACUTE RHINOSINUSITIS
Differential diagnosis of acute rhinosinusitis to exclude on
history and examination:
Dental pain (particularly in cases of unilateral facial pain)
Neuralgic (atypical) facial pain
Temporomandibular joint pain
Migraine
Trigeminal neuralgia
Temporal arteritis
Neoplastic conditions
10MINUTE CONSULTATION
Adult acute rhinosinusitis
J Bird,
1
TC Biggs,
1
M Thomas,
2
RJ Salib
1 3
1
Department of Otolaryngology/
Head & Neck Surgery, University
Hospital Southampton NHS
Foundation Trust, Southampton
SO16 6YD, UK
l
University of Southampton,
Aldermoor Health Centre,
Southampton
!
Clinical Experimental Sciences
Academic Unit, Faculty of Medicine,
University of Southampton,
Southampton
Correspondence to: J Bird
jonathan.bird3!@googlemail.com
doi: 1u.11!6/bmj.fl637
This is part of a series of occasional
articles on common problems in
primary care. The BMJ welcomes
contributions from GPs.
bmj.com
this series
OHearing loss in adults
(BMJ lu1!;!/6:fl/96)
OAssessment and
management of renal colic
(BMJ lu1l;!/6:f93')
OVasectomy
(BMJ lu1!;!/6:f167/)
ODry eye
(BMJ lu1l;!/':e7'!!)
OMinor incised
traumatic laceration
(BMJ lu1l;!/':e63l/)
BMJ | 25 MAY 2013 | VOLUME 346 39
PRACTICE
rhinitis (rhinitis medicamentosa). Antihistamines do
not have a role in the treatment of acute rhinosinusitis.
Steam inhalations have not shown consistent beneht in
the treatment of acute rhinosinusitis, but some clinical
trials have shown symptomatic relief. Signs of impending
complications (orbital, intracranial, etc) should prompt
immediate specialist referral.
In summary, most cases of uncomplicated acute rhi-
nosinusitis will settle with conservative measures. Gen-
eral practitioners should advise patients to keep well
hydrated and to use analgesia and nasal or systemic
decongestants. Topical nasal steroids and nasal saline
douches can be trialled if a patient wishes. In systemically
unwell patients whose symptoms have increased aher hve
days or persisted aher 10 days, a hve day trial of antibiot-
ics in conjunction with nasal decongestants or douches
and topical steroids should be considered. If there is no
improvement aher having taken antibiotics for 48 hours
or if signs of impending complications (orbital or intrac-
ranial) have developed, a referral to ear, nose, and throat
(ENT) services (urgent in the latter scenario) should be
made. For a more detailed management approach, refer
to the algorithm (hgure).
When to refer immediately
The sinuses are in close anatomical proximity to the orbits
and brain. As such, acute rhinosinusitis can lead to com-
plications associated with substantial morbidity or even
mortality. Such complications are thankfully rare but are
more commonly encountered in the paediatric popula-
tion. Immediate referral (same day) for urgent investiga-
tion and intervention should therefore be considered in
the following circumstances:

Periorbital oedema or cellulitis

Displaced globe

Double vision

Ophthalmoplegia

Reduced visual acuity

Frontal swelling

Signs of meningitis or focal neurological signs.
Competing interests: We have read and understood the BMJ Group policy
on declaration of interests and have no relevant interests to declare.
Provenance and peer review: Not commissioned; externally peer
reviewed.
Accepted: December
of 6-15%, although it is ohen self managed without medi-
cal care being sought. The primary cause of acute rhino-
sinusitis is viral, with only 0.5-2% developing secondary
acute bacterial rhinosinusitis. Acute rhinosinusitis is usu-
ally self limiting, as evidenced in most randomised trials.
Therefore, antibiotics should not be routinely prescribed
as they are unlikely to aect the duration or severity of ill-
ness. Antibiotic therapy should be reserved for patients
with progressive or worsening symptoms or for those
who are systemically unwell (for example, fever >38C or
worsening facial pain), with a 5-7 day course being most
appropriate. Amoxicillin-clavulanate (rather than amoxi-
cillin alone because of emergence of antimicrobial resist-
ance among respiratory pathogens) is recommended as
empirical therapy for non-penicillin allergic adults at a dose
of 500 mg/125 mg orally three times daily. In penicillin-
allergic patients, doxycycline (100 mg orally twice daily)
or c larithromycin (500 mg orally twice daily) would be
reasonable choices.
Additional treatment with a systemic decongestant
(such as pseudoephedrine) or topical nasal decongest-
ant (such as xylometazoline) or nasal saline douche (such
as Sterimar or Sinus Rinse) may have modest benehts,
and patients can be advised to purchase these over the
counter if necessary. Intranasal corticosteroids in short
courses (such as mometasone furoate 50 g nasal spray
twice daily for 7-14 days) are eective as monotherapy
for moderate disease and in combination with antibiot-
ics for severe disease. Nasal decongestants should not be
used for more than 10 days as they can induce rebound
USEFUL READING
Thomas M, Yawn BP, Price D, Lund V, Mullol J, Fokkens W,
et al. EPOS primary care guidelines: European position
paper on the primary care diagnosis and management of
rhinosinusitis and nasal polyps a summary. Prim
Care Respir J ;:-.
Fokkens WJ, Lund VJ, Mullol J, Bachert C, Alobid I, Baroody
F, et al. European position paper on rhinosinusitis and
nasal polyps . Rhinol Suppl ;(): p preceding
table of contents, -.
Fokkens WJ, Lund VJ, Mullol J, Bachert C, Alobid I,
Baroody F, et al. EPOS European position paper
on rhinosinusitis and nasal polyps . A summary for
otorhinolaryngologists. Rhinology ;():-
Diagnose acute rhinosinusitis
Acute onset of at least two symptoms, one of which must be
Nasal blockage or purulent nasal discharge
Other symptoms are
Facial pain
Smell disturbance
Perform anterior rhinoscopy for signs
Imaging not required unless complications
Increased symptoms aer days or
persistent symptoms aer days
Symptoms < days
Complications
(such as periorbital
cellulitis, frontal
swelling, double
vision, etc)
day trial of
antibiotics, nasal
decongestant,
and nasal saline
douche, and possibly
topical nasal steroids
Consider day trial of
antibiotics, nasal
decongestant, and
nasal douche
No complications
Refer to ENT Urgent referral to ENT
No improvement
aer hours
Refer to ENT
No improvement aer
hours or progression
in symptoms
Improvement
aer hours
Worsening or
severe symptoms
Continue treatment
for - days
Topical nasal steroids Symptomatic relief
Common cold
(acute viral rhinosinusitis)
Systemically well Systemically unwell
Management algorithm for adults with acute rhinosinusitis
40 BMJ | 25 MAY 2013 | VOLUME 346
ENDGAMES
We welcome contributions that would help doctors with postgraduate examinations
OSee bmj.com/endgames for details
An 81 year old man was referred to a
respiratory doctor with a two year history of
worsening shortness of breath on exertion. He
also described presyncopal symptoms caused
by exertion and one episode of collapse after
walking up a hill.
His medical history included benign
prostatic hypertrophy and encephalitis. He
was a non-smoker, drank little alcohol, and
had previously led an active life.
On examination he was afebrile, his blood
pressure was 13//8/ mm Hg, his heart rate
was 8u beats/min in sinus rhythm, and his
oxygen saturation on room air was 95%.
His jugular venous pressure was not raised.
He had a loud P2 component of the second
heart sound and no ankle oedema. His
lung fields were clear. Electrocardiography
showed ST depression and T wave inversion
anterolaterally. Chest radiography showed
enlarged hila.
An echocardiogram showed normal left
ventricular function with a dilated right
ventricle and pulmonary artery systolic
pressure of 63-68 mm Hg. Computed
tomography pulmonary angiography showed
a dilated right ventricle, dilation of the
pulmonary trunk, multiple webs in pulmonary
artery branches, and thrombus layering along
the right main pulmonary and right upper lobe
pulmonary artery. He was referred to a regional
specialist centre to confirm the diagnosis and
decide further management.
1 What is the diagnosis?
2 What investigations would you perform in
this patient?
3 How should the patient be managed?
/ What are the options for long term
management?
Submitted by Stella Nikolaou and David P Jenkins
A 56 year old man presented with an
injury to the radial aspect of his right wrist
(figure). While undertaking mechanical
repairs to his car the previous day his
metal wrist watch had made contact
with part of the electrical system. He felt
immediate pain but did not implement
any first aid measures or seek medical
attention at the time of injury. Subsequent
skin changes and discomfort at the
periphery of the lesion prompted him
to visit his general practitioner, who
suspected an electrical burn and referred
him to the local burns unit for assessment
and management. He had no medical
history of note. On examination the centre
of the lesion had a leathery appearance,
did not blanch on pressure, and was
insensate.
FOLLOW ENDGAMES ON TWITTER
@BMJEndgames
FOR SHORT ANSWERS See p 23
FOR LONG ANSWERS
Go to the Education channel on bmj.com
The effectiveness of topical chloramphenicol
in preventing wound infection after minor
dermatological surgery was evaluated.
Researchers performed a randomised
placebo controlled double blind superiority
trial. The intervention was a single
application of topical chloramphenicol
ointment to the sutured wound immediately
after suturing. The placebo treatment was a
single application of paraffin ointment. In
total, 972 minor surgery patients with high
risk sutured wounds were recruited. Trial
participants were randomised to topical
chloramphenicol ointment (n=/88) or
placebo (n=/8/). The primary outcome was
infection on the agreed day of removal of
sutures or sooner if the patient re-presented
with a perceived infection.
The critical level of significance for
statistical testing was set at u.u5 (5%).
The proportion of participants with an
infection was significantly lower in the
chloramphenicol group than in the placebo
group (6.6% v 11.u%; difference /./%,
95% confidence interval 7.9% to u.8%;
P=u.u1u). The researchers concluded that
the application of a single dose of topical
chloramphenicol to high risk sutured wounds
after minor surgery produced a statistically
significant yet moderate absolute reduction
in infection rate.
Which of the following statements,
if any, are true?
a) The P value provides a direct statement
about the size of the difference between
groups in the proportion of patients with
wound infection
b) The P value provides a direct statement
about the direction of the difference
between groups in the proportion of
patients with wound infection
c) The P value provides a dichotomous test of
significance of the statistical hypotheses
d) The 95% confidence interval provides a
dichotomous test of significance of the
statistical hypotheses
Submitted by Philip Sedgwick
STATISTICAL QUESTION
P values or confidence intervals?
PICTURE QUIZ An unusual burn
CASE REPORT
A treatable cause of shortness of breath
Submitted by C J Mitchell, Z Ahmad, and M S Khan
1 From the photograph, estimate the total body surface area and depth of this burn.
2 How would you manage a patient presenting with an electrical burn?
3 What complications can occur with an electrical burn?
BMJ | 25 MAY 2013 | VOLUME 346 41
LAST WORDS
The NHS is no more
important to health
and welfare than
policing and social
care, and it is less
important than
income inequalities
and welfare reforms
The NHS is in a comfort zone: com-
placent, defensive, and out of touch
with the economic realities. The NHS is
no more important to health and wel-
fare than policing and social care, and it
is less important than income inequali-
ties
5
and welfare reforms. It is awash
with inemciencies; duplication; rigid
boundaries; and unnecessary referrals,
investigations, and treatments; and it is
plagued by poor communication and
tribalism.
And these endemic problems are mir-
rored throughout the developed world.
There can be no real reform of health-
care without the economic leverage of
budgetary cuts.
Doing less, working dierently, and
thinking more is the best of medicine.
Cuts are inevitable: time to get on with
them, and make a better NHS. See you
all at Lidl.
Des Spence is a general practitioner, Glasgow
destwo@yahoo.co.uk
I met a friend in Lidl. We shook our
heads. I said, They constantly com-
plain, bringing their own bags to con-
ceal the fact theyre shopping at Lidl,
and they dont follow the queuing
etiquette. Bloody Marks and Spencer
types! The middle classes are moving
to Lidl, taking our trolleys and buying
up all the risotto rice. Next, middle
class children will be getting a bus pass
rather than a car on their 18th birth-
day, and spend their gap year working
in Primark in Chelmsford. This is the
economic reality, and it is delivering a
social correction. And this is why NHS
funding should not be ringfenced.
This budgetary protection is politi-
cally motivated (always a bad premise),
with a Conservative government unwill-
ing to be seen to cut NHS spending. But
the truth is that the Labour govern-
ment invested in the NHS too rapidly
and without enough thought. Its new
contracts for general practitioners, con-
sultants, and private hnance initiatives
were hugely expensive and disastrously
inemcient. And there has thus far been
only a phony debate about NHS cuts,
1

with the same unchallenged, simplistic
assumptions.
Firstly, that health ination always
outstrips normal ination. This is not
so: health ination has dipped hugely
in the United States under current eco-
nomic pressures.
2
We must learn to be
sceptical of new unproved drugs, to not
be seduced by new equipment and to
stop the mindless escalation of unnec-
essary investigations and interventions.
Secondly, we need to remember that
larger health spending as a proportion
of GDP has limited eect on outcomes:
more is not necessarily better.
3
Lastly,
there is an assertion that an ageing
population will put more pressure on
the NHS. But health costs are linked to
morbidity, not age. And no one seems
to have noticed the decline in ischae-
mic heart disease, stroke, and lung
diseasechronic disease is in terminal
decline.
4
There is no demographic time
bomb.
What would you not give, when
pain has you in its grip, to know the
joy of the normal? What indeed?
Thermogene, hrst marketed in the
1900s by the Belgian pharmacist
Charles Vandenbroeck,
1
claimed to
be the good Genie of Bliss to the
pain-racked suerera priceless boon
when pain gnaws and stabs.
Thermogene was cottonwool
impregnated with chilli sauce. The
active ingredient in chilli sauce is
capsaicin. Capsaicin causes local
burning pain and erythema and may
indeed act as a counterirritant. A
capsaicin cream is licensed for the
treatment of osteoarthritic pain and
a rather stronger cream for neuralgic
pain; they are at best only moderately
eective.
2

3
But there is more to capsaicin than
counterirritation. Ninety years aher
Vandenbroeck launched his product,
a specihc receptor for capsaicin was
than inicting more pain on the
unfortunate suerer.
The logic in using shock doses of
capsaicin is that it might, applied to
cutaneous nerves briey, desensitise
nerve endings for much longer.
7

8
A
capsaicin 8% patch, applied for half
an hour or so, has been claimed to
relieve neuropathic pain for up to
three months, though a Cochrane
review is lukewarm.
9
Capsaicin 8%
improved pain by 50% or more in
only one in 12, and as expected
it ohen caused local pain that
sometimes lasted for days.
So unless you have a burning desire
for such treatments, it may be best to
take the therapeutic claims for chilli
with a tiny pinch of salt.
Robin Ferner is director, West Midlands Centre
for Adverse Drug Reactions, Birmingham
R.E.Ferner@bham.ac.uk
The logic in using
shock doses of
capsaicin is that it
might, applied to
cutaneous nerves
briefly, desensitise
nerve endings for
much longer
identihed. The transient receptor
potential channel, vanilloid subfamily
member 1 (TRPV1) receptor
responds to pure capsaicin. Thorough
investigators also observed responses
to habaero, Thai green, yellow wax,
and poblano verde peppers.
4
As chilli
ahcionados know, dierent varieties
cause dierent degrees of pain. The
taste for chilli also seems to require
habituation, as those who have
enjoyed mild food in chilli-eating
parts of the world know.
Very high concentrations of
topically applied capsaicin have been
advocated to treat neuropathic pain.
High doses of capsaicin are expected
to cause burning and erythema.
Indeed, capsaicin induced pain is
a convenient experimental model
for pain mediated via those TRPV1
receptors, so it would be logical
to try to discover TRPV1 receptor
antagonists to prevent pain, rather
FROM THE FRONTLINE Des Spence
Stop protecting the NHS budget
DRUG TALES AND OTHER STORIES Robin Ferner
A cool response to a hot substance
Twitter
Follow Des Spence on
Twitter @des_spence
42 BMJ | 25 MAY 2013 | VOLUME 346

MINERVA
Send comments or suggest ideas to Minerva: minerva@bmj.com
An unusual burn
Try the picture quiz in
ENDGAMES, p
Minerva finds it outrageous that the results of many
clinical trials are never published or published in
such a way that false conclusions can be drawn.
Publication bias is so common that we have come
to accept it as a fact of life, but it is deeply unethical
and fraught with potential harm to patients. In 2uu2,
three authors completed a meta-analysis of the
published literature on the effects of erythropoiesis
stimulating agents in patients with cancer. In Swiss
Medical Weekly they revisit their conclusions in the
light of information that was available in 2uu2 but
unpublished (2u13;1/3:w13776; doi:1u.//1//
smw.2u13.13776). The overall survival benefit
they reported in their first analysis disappeared.
If the published data are so incomplete that we
cannot use them for a meta-analysis how can we
possibly use them to make healthcare decisions
for patients? they ask, adding that Unrestricted
access to clinical study protocols including
amendments, to clinical study reports as well
as to IPD [individual patient data] is needed to
ensure timely detection of both beneficial and
harmful effects of healthcare interventions.
Minerva strongly agrees and hopes that this will be
incorporated into the forthcoming revision of the
Helsinki Declaration.
Lyme disease is fairly ubiquitous in many parts of
the temperate world, particularly in New England,
where John Updike once wrote an elegiac short
story about population drift from a small town
caused by Borrelia burgdorferi. A vaccine against
outer surface proteins carried by several Borrelia
spp is currently being trialled in Austria and
Germany, and an article in The Lancet Infectious
Diseases describes a promising level of antibody
production from an adjuvanted form of the vaccine
(2u13; doi:1u.1u16/S1/73-3u99(13)7u11u-5).
Further studies are needed, but if the vaccine
becomes available, mournful small towns in New
England may begin to fill up again; though when
Minerva last visited Old Lyme, the place seemed to
be doing very nicely already.
Advances in the drug treatment of heart failure
have largely stalled over the past decade, but
plenty can be done to improve the outlook and
quality of life of patients by better use of existing
treatments. A follow-up study of the REVERSE
cohort in European Heart Journal shows that,
in patients with mild systolic heart failure with
conduction delay, resynchronisation therapy
(biventricular pacing) corrects adverse ventricular
remodelling and is associated with low mortality
over five years (2u13 doi:1u.1u93/eurheartj/
eht16u). Whats more, patients felt better over
this period.
Early in your medical career you quickly learn
that people have different chronotypes. Student
companions who seemed perfectly normal
suddenly morph into mad surgeons capable of
operating night and day, while others soon come to
wear the wilted look of a junior doctor chronically
deprived of sleep. Sustained by her companion
owl, and made of immortal fabric, Minerva herself
never sleeps, but was interested to read about
the developing science of chronotyping in the
Journal of Biological Rhythms (2u13;28:1/1-51;
doi:1u.1177/u7/873u/12/75u/2). A total of
238 shift workers were chronotyped with the
Munich chronotype questionnaire for shift workers,
which collects information about shift dependent
sleep duration and sleep timing. There are wide
variations in effect, and Minerva feels it would be
a good idea if medical students were chronotyped
on entry and encouraged towards the specialty that
best suits their need for sleep.
Every doctor should know how to diagnose cluster
headache, an agonising unilateral syndrome of
pain, usually close to the eye and accompanied
by autonomic symptoms in the same area. It is
most common in young men. A recent article in
Cephalalgia, a journal devoted to the study of
head pain syndromes, describes two stepwise
questionnaires for diagnosing cluster headache
the Leiden University cluster headache analysis
program (LUCA) and its shorter version, QATCH
(2u13; doi:1u.1177/u3331u2/13/79835).
These may come in handy for borderline cases,
but the authors conclude that men with severe
unilateral headaches of short duration occurring
at long intervals (months or years) are very likely
to have cluster headache.
One of the most remarkable developments in the
history of European culture was the craze for courtly
love, which started in the south of France late in the
11th century and spread northwards through the
passionate songs of the troubadours. The literary
scholar CS Lewis described it as love of a highly
specialized sort, whose characteristics may be
enumerated as humility, courtesy, adultery, and the
religion of love. The women of France still prefer
those who court them to show signs of musical
ability, although from a study in Psychology of
Music it seems that they have become less choosy
(2u13; doi:1u.1177/u3u5735613/82u25).
Three hundred young women were solicited in
the street for their phone number by a young male
confederate who held either a guitar case or a
sports bag in his hands or had no bag at all. Results
showed that holding a guitar case was associated
with greater compliance to the request, thus
suggesting that musical practice is associated with
sexual selection. The next trial will have them sing
O rosa bella beneath the castle window at dawn.
For 1u years the Klevis Kola Foundation (www.
kleviskola.org/), an organisation founded by
medical students at St Georges Hospital, has
been helping refugee and asylum seeker families
in south London. With the retreat of social services
and the growing hostility towards immigrants
their work is expanding. In the autumn they will
celebrate 1u years of work and hope that former
students will join them. Contact Sarah Sibley
atsarahsibley@kleviskola.org.
This 8/ year old woman had a sigmoid
adenocarcinoma resected by an emergency
Hartmanns procedure in 2uu6, after which
she developed a large parastomal hernia. She
had chosen not to receive regular support for
stoma care in the community and had been self
managing her stoma for the past six years. Her
five year surveillance had shown no recurrence
of disease. The unusual appearance of her
stoma aperture was initially thought to be due
to recurrence of her cancer, but was actually
caused by ill fitting stoma bags, effluent
leakage, and subsequent overhealing of the
damaged mucosal surface, resulting in multiple
granulomata. These were managed with silver
nitrate and appropriately sized stoma bags.
Madeha Khan (madeha.khan@gmail.com), foundation
doctor, Rebekah Schiff, consultant geriatrician,
Department of Ageing and Health, St Thomas Hospital,
London SE EH, UK
Patient consent obtained.

BMJ - 25 May 2013

Încărcat de

Informații document

Descriere originală:

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

BMJ - 25 May 2013

Încărcat de

Drepturi de autor:

Formate disponibile

THIS WEEK

BMJ | 25 MAY 2013 | VOLUME 346

25 May 2013 Vol 346

The bill was introduced in parliament in the

Seddon wrote: If we want the NHS to

Seddon has acknowledged that some hos-

more than four times the number killed in combat.

Life expectancy of people with mental illness compared with

Anoop Dinesh Shah,

STUDY QUESTION How do multiple linked data sources

London School of Hygiene and

Department of Epidemiology and

Clinical Practice Research Datalink

Utrecht Institute for Pharmaceutical

Barts and the London School of

Bariatric surgery is recommended if all of the following criteria are fulfilled:

It becomes clear, however, that

and carbohydrates are not

National Collaborating Centre for

Peterborough and Stamford

and .% in secondary care.

S-ar putea să vă placă și